DIVERSITY AND ECOLOGY OF LACUSTRINE HAPTOPHYTE ALGAE: IMPLICATIONS FOR PALEOTHERMOMETRY

By:

SUSANNA M. THEROUX

B.S. Williams College, Williamstown, MA, 2005

M.Sc. Brown University, Providence RI, 2009

A dissertation submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy in The Department of Geological Sciences at Brown University

PROVIDENCE, RHODE ISLAND

MAY 2013

© Copyright 2013 by SUSANNA M. THEROUX

ii This dissertation by SUSANNA M. THEROUX is accepted in its present form by The

Department of Geological Sciences as satisfying the dissertation requirement for the degree of

Doctor of Philosophy.

Date______

______

Yongsong Huang, Ph.D., Co-Advisor

Date______

______

Linda Amaral-Zettler, Ph.D., Co-Advisor

Recommended to the Graduate Council

Date______

______

Timothy D. Herbert, Ph.D., Reader

Date______

______

James M. Russell, Ph.D., Reader

Date______

______

Marco J.L. Coolen, Ph.D., Reader

Approved by the Graduate Council

Date______

______

Peter Weber, Dean of the Graduate School

CURRICULUM VITAE

SUSANNA M. THEROUX

EDUCATION

Ph.D.: Geological Sciences, Brown University, Providence, RI, Brown-Marine Biological Laboratory in Woods Hole graduate program (2012 expected)

M.Sc.: Geological Sciences, Brown University, Providence, RI, Brown-Marine Biological Laboratory in Woods Hole graduate program, 2009

B.A.: Biology and Geology with honors, Williams College, Williamstown, MA, 2005

FELLOWSHIPS & AWARDS

- American Association of University Women (AAUW) Dissertation Fellowship, 2011 - Geological Society of America (GSA) Student Research Grant, 2010 - Brown-MBL Graduate Fellowship in Biological and Environmental Sciences, 2007- 2009 - Professor Jack L. Strominger Graduate Fellowship, Brown University, 2008 - Freeman Foot Award (for thesis presentation), Williams College, 2005 - 1960s Scholar in Geosciences, Williams College, 2005 - Student Scholar Athlete, Williams College, 2005 - Dean’s List, Williams College, 2001-2005

TEACHING EXPERIENCE

- Woods Hole Semester in Environmental Studies, 2011, student mentor - Woods Hole Partnership Education Program, 2011, guest lecturer and mentor - GEO 031: Paleoecology and the Fossil Record, Brown University, Teaching Assistant and Lab Instructor - GEO 020: Introduction to Oceanography, Brown University, Teaching Assistant

PUBLICATIONS

Theroux, S., Toney, J.L., Amaral-Zettler, L., Huang, Y. in prep. Production and temperature sensitivity of long chain alkenones in cultures of Pseudoisochrysis paradoxa, nomen nudum. Geochimica et Cosmochimica Acta.

Theroux, S., Huang, Y., Amaral-Zettler. 2012. Comparative molecular microbial ecology of spring haptophyte blooms in an arctic oligosaline lake in Greenland. Frontiers in Microbiology. doi: 10.3389/fmicb.2012.00415.

iv Toney, J., Theroux, S., Andersen, R., Coleman, A., Amaral-Zettler, L., Huang, Y. 2012.

Culturing of the first 37:4 predominant lacustrine haptophyte: Geochemical, biochemical, and genetic implications. Geochimica et Cosmochimica Acta 78, 51-64.

Theroux, S., D’Andrea, W.J., Toney, J., Amaral-Zettler, L.A., Huang, Y. 2010. Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes: Implications for continental paleotemperature reconstructions. EPSL 300, 311-320.

Amaral-Zettler,L., Zettler, E., Theroux, S., Palacios, C., Aguilera, A., Amils, R. 2010. Microbial community structure across the tree of life in the extreme Río Tinto. ISME Journal. doi:10.1038/ismej.2010.101

Stoll, H. M., Ziveri, P., Shimizu, N., Conte, M., Theroux, S. 2007. Relationship between coccolith Sr/Ca ratios and coccolithophore production and export in the Arabian Sea and Sargasso Sea. Deep Sea Research II: 54, 581-600.

Theroux, S. 2005. Effects of nutrient limitation on the productivity of coccolithophore algae and the paleoclimatic implications. BA thesis, Williams College.

PRESENTATIONS & POSTERS

Theroux, S., Diversity and ecology of Alkenone-producing haptophyte algae. Invited. Lamont Doherty Earth Observatory Spring Lecture Series. April, 2012.

Theroux, S., Toney, J., Andersen, R., Bohn, R., Nyren, P., Amaral-Zettler, L., Huang, Y. Dynamics of a haptophyte bloom in Lake George, ND: Implications for alkenone- based temperature reconstructions. American Geophysical Union Fall Meeting, 2011.

Salacup, J.M., Theroux, S., Herbert, T.H., and Prell, W.L. Time-series of water column k' alkenones and 18S rRNA confirm that U37 is a viable SST proxy in Narragansett Bay, RI. American Geophysical Union Fall Meeting, 2011.

Theroux, S., Amaral-Zettler, L., Huang, Y. Insights into haptophyte bloom dynamics in an arctic lake using massively parallel tag sequencing. International Society of Microbial Ecology 13, 2010.

Theroux, S., D'Andrea, W. J., Toney, J.L., Amaral-Zettler, L., Huang, Y. Relating Phylogeny to Alkenone Distributions in Lacustrine Alkenone-Producing Haptophytes: Implications for Continental Paleotemperature Reconstructions. American Geophysical Union Fall Meeting, 2008.

Toney, J.L., Fritz, S., Baker, P.A., Grimm, E.C., Nyren, P., Theroux, S., Huang, Y. Environmental and Climatic Control on the Occurrence and Abundance of Long Chain

v Alkenones in Lakes of the Interior United States. American Geophysical Union Fall Meeting, 2008.

Zettler, E., Amils, R., Theroux, S., Palacios, C., Sogin, M., Amaral-Zettler, L. Protistan distribution in relation to spatial variations of extreme geochemical parameters in the Rio Tinto, Spain. Phycology Society of American Annual Meeting, 2007.

Theroux, S., Burke, A., Stoll, H., Shimuzu, N., Ziveri, P. Constraining nutrient effects in monsoon-driven productivity shifts during the Mediterranean sapropel events. Presentation, Geological Society of America Northeastern Section annual meeting, 2005.

EMPLOYMENT

- Research Assistant, 2005-2007 Marine Biological Laboratory, Woods Hole, MA Josephine Bay Paul Center for Comparative Molecular Biology and Evolution, Dr. Linda Amaral-Zettler Laboratory

- Summer Research Fellowship in Geosciences, 2004-2005 Williams College, Williamstown, MA Dr. Heather Stoll Laboratory

- Summer Research Fellowship in Geosciences, 2003 Vrije Universiteit, Amsterdam, NL Dr. Patrizia Ziveri Lab oratory

MEMBERSHIPS & CERTIFICATIONS

Algal Biomass Organization (ABO) American Association of University Women (AAUW) Earth Science Women’s Network (ESWN) American Geophysical Union (AGU) Sigma XI Scientific Research Society Geological Society of America (GSA) International Society for Microbial Ecology (ISME) Phycological Soceity of America (PSA) Spanish Language Advanced Proficiency

FIELD EXPERIENCE

Lake George, North Dakota: lake sampling, algal culturing, 2011 Kangerlussuaq, Greenland: lake sampling, algal culturing, 2009 Kangerlussuaq, Greenland: lake sampling, coring expedition, algal culturing, 2008 Kangerlussuaq, Greenland: lake sampling, lake sediment coring, 2007 Tiputini Biodiversity Station, Parque Nacional Yasuni, Ecuador: 2004

vi

Dedicated to my family

vii ACKNOWLEDGEMENTS

I would like to extend my greatest appreciation to my co-advisors, Linda Amaral-Zettler and Yongsong Huang. I would also like to thank the members of my thesis committee,

Timothy Herbert, James Russell, Zoe Cardon and Marco Coolen. I would like to thank the Earth Systems History group at Brown University and the Josephine Bay Paul Center at the Marine Biological Laboratory. I would also like to thank my friends and family, especially Mom, Dad, Katie, Katie and Casper.

viii Table of Contents CURRICULUM VITAE ...... iv ACKNOWLEDGEMENTS ...... viii CHAPTER 1 ...... 1 1. Introduction ...... 2 1.1 Alkenone production in haptophyte algae ...... 2 1.2 Alkenone-based paleothermometry ...... 3 1.3 Molecular biological approaches to temperature reconstructions ...... 5 1.4 Dissertation outline ...... 6 References ...... 10 CHAPTER 2 ...... 13 Abstract ...... 14 2. Methods ...... 17 3. Results ...... 23 4. Discussion ...... 27 5. Conclusions ...... 32 CHAPTER 3 ...... 50 Abstract ...... 51 1. Introduction ...... 53 2. Materials and Methods ...... 55 3. Results and Discussion ...... 61 4. Conclusions ...... 67 Acknowledgements ...... 68 References ...... 69 CHAPTER 4 ...... 84 Abstract ...... 86 Acknowledgements ...... 96 References ...... 97 CHAPTER 5 ...... 108 Abstract ...... 109 1. Introduction ...... 110 2. Methods ...... 112

ix 3. Results ...... 116 4. Discussion ...... 119 5. Conclusions ...... 122 Acknowledgements ...... 123 References ...... 125 CHAPTER 6 ...... 141 Abstract ...... 142 1. Introduction ...... 143 3. Results ...... 148 4. Discussion ...... 151 5. Conclusions ...... 154 Acknowledgements ...... 155 References ...... 156 CHAPTER 7 ...... 167 Abstract ...... 168 1. Introduction ...... 169 2. Methods ...... 171 3. Results ...... 178 4. Discussion ...... 184 5. Conclusions ...... 189 Acknowledgements ...... 190 References ...... 192 Supplemental ...... 209 CHAPTER 8 ...... 213 1. Conclusions ...... 214

x LIST OF TABLES CHAPTER 2 Table 1. Lake locations, maximum depth and salinity...... 41 Table 2. Primers used for SSU rRNA gene amplification...... 42 Table 3. Percent OTU representation in lake sediment clone libraries...... 43 Table 4: Uncorrected distance matrices of pairwise comparisons among sequences...... 44 Table 5. Alkenone trends among haptophyte algae...... 45 CHAPTER 3

Table 1. BrayaSø water column C37 alkenone concentration and haptophyte rRNA gene copy number...... 75 Table 2. Sequencing summary, OTU distributions, and diversity estimates...... 76 Table 3. Most abundant bacterial and eukaryotic OTUs...... 77 CHAPTER 4 Table 1. Average growth rate, cell concentrations and alkenone concentration of Pseudoisochrysis paradoxa...... 101 CHAPTER 5 Table 1. Environmental characteristics of lakes in this study...... 129 Table 2. OTU distribution among the six samples by percentage...... 130 Table 3. Salinity reconstruction using Liu et al. (2008) paleosalinity proxy based on

%C37:4...... 131 CHAPTER 6 Table 1. Alkenone abundance data for < 3µm culture experiment...... 160 CHAPTER 7 Table 1. Haptophyte Hap-B #903 culture cell counts and alkenone concentrations. .... 195 Table S1. Hap-B culture 18S rRNA partial gene sequence...... 209 Table S2. Most abundant eukaryote OTU Ion Torrent sequence reads...... 210

xi LIST OF FIGURES

CHAPTER 2 Figure 1. Map of alkenone-containing lakes...... 46 Figure 2. A consensus Bayesian phylogenetic tree depicting 18S rRNA gene-inferred relationships among haptophyte algae...... 47 Figure 3. A consensus Bayesian phylogenetic tree depicting 18S rRNA gene-inferred relationships among haptophyte algae using only full-length sequences...... 48 Figure 4. Lake sediment alkenone signatures...... 49 CHAPTER 3 Figure 1. Site map showing Lake BrayaSø in the Kangerlussuaq region, Greenland.. .. 80 Figure 2. BrayaSø water column...... 81 Figure 3. (A) June average air temperatures for 2007 and 2009. (B) Average monthly air temperatures for 2007 and 2009...... 82 Figure 4. Venn diagrams of microbial communities...... 83 CHAPTER 4 Figure 1. Alkenone cell concentration versus growth rate in P. paradoxa cultures. .... 102 k Figure 2. (Left panel) Pseudoisochrysis paradoxa U37 calibration (Right panel) k' Pseudoisochrysis paradoxa U37 calibration...... 103 Figure 3. Gas chromatogram of P. paradoxa cultures at 10°C. Inset: Photomicrograph of P. paradoxa culture. Scale bar 5µm...... 104 Figure 4. Average alkenone concentrations for P. paradoxa cultures at different temperatures...... 105 k Figure 5. Comparison of P. paradoxa U37 calibration with other species and lake-based calibrations...... 106 k' Figure 6. Comparison of P. paradoxa U37 calibration with other species and lake-based calibrations...... 107 CHAPTER 5 Figure 1. Map of lake locations in the Qinghai basin. Inset of Tibetan plateau with rectangle indicating the sampling locations ...... 132 Figure 2. Water column profiles of sampling sites Qinghai 2 and Qinghai 12...... 133 Figure 3. A consensus Bayesian phylogenetic tree depicting 18S rRNA gene-inferred relationships among haptophyte algae...... 135 Figure 4. Concentrations and percent abundances of alkenone homologs for each lake in

xii this study...... 136 Figure 5. Non-metric multidimensional scaling plot of China lake samples and their Bray-Curtis similarity in haptophyte OTU distributions...... 137 Figure 6. Cluster plot of haptophyte OTU distributions among lakes in this study based on Bray-Curtis similarity...... 138

Figure 7. Relationship between lake water sample salinities and percent C37:4 alkenones in lakes from this study...... 139 Figure S1. Relationship between lake salinity and water temperature for lakes in this

study. %C37:4 concentrations as shown...... 140 CHAPTER 6 Figure 1. Gas chromatograms of culture size fractions...... 161 Figure 2. Percent abundances of < 3µm culture alkenones at all temperatures...... 162 Figure 3. Abundance of C37 isomers in the < 3µm cultures...... 163 Figure 4. Alkenone unsaturation indices for the < 3µm fraction culture...... 164 Figure 5. Lake George enrichment culture < 3µm fraction tagged with haptophyte and Lake George Hap-A specific probes...... 165 Figure 6. Photomicrograph of < 3µm culture...... 166 CHAPTER 7 Figure 1. Map of North Dakota (inset) and Lake George, ND...... 196 Figure 2. Lake George water column throughout the sampling period at 0m, 5m, and 10m depths...... 197 Figure 3. Nutrient concentrations during the Lake George seasonal cycle...... 198 Figure 4. Percent concentrations of alkenone isomers and Ion Torrent sequence read numbers of Hap-A and Hap-B throughout the seasonal cycle at 0m, 5m, and 10m depths...... 199 k Figure 5. A. U37 calibrations for Lake George 2011 in situ calibration, Hap-B isolate #903, and Hap-A <3µm culture...... 200 Figure 6. qPCR haptophyte gene copy number concentrations throughout the course of the Lake George seasonal cycle...... 202 Figure 7. Haptophyte phyla abundances from Ion Torrent OTUs at 0m, 5m, and 10m depths...... 203 Figure 8. A. Cluster analysis of Lake George Ion Torrent microbial community

xiii similarity among samples...... 204 Figure 9. A. Canonical correspondence analysis of Lake George geochemical gradients and OTU abundances...... 206 Figure 10. Hap-B strain #903 alkenones and morphology...... 207 Figure 11. Mock bloom event alkenone signatures and haptophyte DNA sequence taxonomy...... 208 k Figure S1. U37 calibration using only samples from 0m, 5m, and 10m depths...... 212

xiv

CHAPTER 1

INTRODUCTION

SUSANNA M. THEROUX

BROWN UNIVERSITY, Department of Geological Sciences MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

1 1. Introduction

In the face of climate change, the need for accurate climate forecasting relies heavily upon accurate climate hindcasting. Molecular fossils preserved in the geologic record present ideal opportunities for reconstructing ancient environments.

However, a carpenter is only as good as his (her) tools and likewise, a paleoclimatologist is only as good as his (her) proxy. In this vein, this thesis seeks to improve our understanding of extant, alkenone-producing haptophyte algae in order to improve our application of the alkenone-based paleotemperature proxy.

1.1 Alkenone production in haptophyte algae

Haptophyte algae are a globally important group of photosynthetic microbes.

Haptophytes are responsible for large-scale bloom events throughout the world’s oceans, are credited as being the largest producers of calcium carbonate on the planet and are arguably the dominant carbon-fixers in the ocean, even out-pacing cyanobacteria (Liu et al., 2009). Certain haptophyte species within the Class

Isochrysidales (Division Haptophyta) produce long chain polyunsaturated alkenones

(LCAs) (Marlowe et al., 1984). These diagenetically stable compounds have been found in sediments dating back to the lower Aptian (120.5 Ma) although the original haptophyte species responsible for these alkenones is unknown (Brassell et al., 2004).

In modern marine environments, alkenone production is dominated by species

Emiliania huxleyi and its close relative Gephyrocapsa oceanica. In the lesser-studied coastal and lacustrine environments, Isochrysis galbana and Chysotila lamellosa are known alkenone-producers (Marlowe et al., 1984).

2 The physiological role of the alkenone molecule is still unknown. Originally believed to be a membrane lipid, this theory was dismissed upon discovery of the alkenone’s trans (versus typical membrane cis) unsaturated bonds (Rechka and

Maxwell, 1988). The observation that alkenones accumulated in cells under nutrient deprivation (Rontani et al., 2004) and are metabolized when cells are incubated in the dark (Eltgroth et al., 2005) suggested they are used as an energy storage molecule.

Observation studies localized the alkenone lipid to cytoplasmic vesicles often associated with chloroplasts (Eltgroth et al., 2005). A proposed biosynthetic pathway theorized that alkenones are synthesized in the chloroplast from an acetyl- or propinoyl-SCoA precursor, exported to the cytoplasm, and progressively unsaturated by a series of desaturases (Rontani et al., 2006).

1.2 Alkenone-based paleothermometry

Only four species of Haptophyta have been systematically cultivated in pure culture to gauge alkenone production: predominant, and bloom-forming, marine haptophyte Emiliania huxleyi (Prahl et al., 1988; Prahl et al., 2003; Conte et al.,

1998), close marine relative Gephyrocapsa oceanica (Sawada et al., 1996; Conte et al., 1998), and coastal/lacustrine species Chrysotila lamellosa (Marlowe et al., 1984;

Sun et al., 2007) and Isochrysis galbana (Marlowe et al., 1984; Versteegh et al.,

2001).

Haptophytes fortuitously record the temperature of their surrounding waters in the degree of unsaturation of their alkenones; as temperature decreases, the alkenones grow systematically more unsaturated. This relationship between water temperature

3 and alkenone unsaturation was first described by Brassell et al. (1986) and

incorporated the relative concentrations of alkenones of 37 carbon atoms (C37) with either two, three of four points of unsaturation (C37:2, C37:3, C37:4, respectively). This

k relationship, the U37 alkenone-based paleotemperature proxy, was defined by

Brassell et al. (1986) as:

k CC: 2 − [C37: 4] U = 37 C37: 2 + C37: 3 + [C37: 4]

k In environments with negligible C37:4 concentrations, the U37 index was modified by

Prahl and Wakeham (1987) to:

k C37: 2 U = 37 C37: 2 + C37: 3

The empirical definition of the linear relationship between the increasingly

unsaturated compounds (C37:3) with cooler water temperatures in cultures of E.

k huxleyi established the first calibration for the U37 proxy to calculate ancient sea surface temperatures (Prahl and Wakeham, 1987):

k U37 = 0.040T - 0.11

This relationship was confirmed with a global coretop calibration by Muller et al.

(1998):

k U37 = 0.033T + 0.044

k These U37 calibrations have been used successfully to reconstruct sea surface temperatures from sediment cores throughout the world oceans.

4 k The desire to extend the use of the U37 paleotemperature proxy to the continents focused attention on lacustrine alkenone archives. In contrast to marine sediments, lake alkenones are characterized by an abundance of tetraunsaturated alkenones, indicating the presence of a different alkenone-producing haptophyte species or an environmental control dictating a distinct alkenone profile. Alkenones have been identified in lakes sediments around the world, including Antarctica,

Europe, Greenland, South America, China and the western United States (Thiel et al.,

1997; Coolen et al., 2004; D'Andrea and Huang, 2005; Liu et al., 2008; Pearson et al., 2008; Toney et al., 2010; Zink et al., 2001). Lacustrine alkenone unsaturation ratios correspond to seasonal surface water temperatures and mean annual air temperature (Chu et al., 2005; D'Andrea and Huang, 2005; Sun et al., 2007; Zink et al., 2001) further confirming the potential utility of alkenone archives from lake

k sediments. However, because of the paucity of lacustrine U37 calibrations, efforts to reconstruct paleotemperatures from a suite of German lakes yielded erroneous alkenone-based paleotemperature estimates (Zink et al., 2001). Clearly, a new method

k of identifying haptophytes species, and their U37 calibrations, is needed.

1.3 Molecular biological approaches to temperature reconstructions

For the past eight years, there has been a movement towards identifying haptophyte species via their most specific biomarker: DNA. A study of preserved

DNA in the sediments of Ace Lake, Antarctica revealed the presence of multiple haptophyte species closely related to known coastal/lacustrine species Isochrysis galbana (Coolen et al., 2004). Like Ace Lake, the sediments of a suite of oligosaline

5 lakes in southwestern Greenland were dominated by the C37:4 alkenone (D’Andrea and

Huang, 2005). However, unlike Ace Lake, the Greenland lake sediments also had the

presence of the C38 methyl ketone, typically only seen in marine alkenone profiles

(D’Andrea and Huang, 2005; Schulz et al., 2000). Haptophyte 18S ribosomal RNA

(rRNA) gene sequences revealed a novel clade of haptophyte algae, one distinct from both the marine and lacustrine haptophyte clades (D’Andrea et al. 2006). These preliminary studies set the stage for the global exploration of alkenone-producing haptophytes in lake environments. Using a multidisciplinary approach, this thesis aims to i) identify and classify novel species of alkenone-producing haptophytes, ii)

k calibrate their U37 alkenone-unsaturation indices and iii) determine how haptophyte bloom events impact the lacustrine alkenone sediment record.

1.4 Dissertation outline

This dissertation applies molecular biological techniques to answer enduring questions in the development of the alkenone-based paleothermometry in lake environments.

Chapter 2 is the first global survey of lacustrine sediments to couple alkenone profiles with haptophyte DNA sequences. We analyzed 18S rRNA gene sequences specifically for haptophytes to test the validity of inferring haptophyte species from alkenone lipid profiles alone. Our results showed that DNA sequences were the more specific identifier of haptophyte species and that lakes with identical alkenone profiles can have different haptophyte communities. Additionally, some lakes hosted multiple, distantly related haptophyte species, indicating that alkenone profiles may

6 be composite profiles with unknown contributions by either species. In this study we also sequenced paleo-DNA and found the conservation of alkenone-producing haptophytes in downcore sediments from Lake BrayaSø, Greenland. This result

k green-lights the use of a single U37 calibration to be applied when reconstructing paleotemperatures from the BrayaSø sediment record.

Chapter 3 is a high-throughput DNA sequencing survey of the microbial communities surrounding a haptophyte bloom. We analyzed water samples from both the 2007 and 2009 haptophyte bloom events in Lake BrayaSø, Greenland, and characterized the bacterial and eukaryotic microbial communities. We found a surprising lack of overlap in microbial taxa between the two bloom years. However, this study confirmed the presence of only one haptophyte species in Lake BrayaSø, a

k necessary condition for future application of a single U37 calibration. Using quantitative PCR, our results also confirmed that alkenone concentrations correspond to haptophyte cell concentrations in the water column. This study serves as a benchmark microbial survey in an arctic oligosaline lake, providing baseline knowledge of arctic microbial communities as they encounter rapid climate change.

Chapter 4 addressed a prerequisite for the further development of the

k k lacustrine U37 proxy: the definition of species-specific U37 calibrations. In this study we grew haptophyte Pseudoisochrysis paradoxa in pure culture at different temperatures to define its alkenone unsaturation—temperature relationship.

Pseudoisochrysis paradoxa is the closest relative to a novel species of haptophyte identified from Lake George, ND, and this study serves as an example of how in situ investigations of alkenone production by haptophytes requires a complimentary

7 dataset of reference haptophyte species that have been examined in silico. The

cultures of P. paradoxa had dominant C37:3 isomers and high concentrations of

k alkenones per cell at the lowest growth rates. The U37 calibration for P. paradoxa was most similar to the calibration for species C. lamellosa instead of P. pardoxa’s closest relative, Isochrysis galbana.

Chapter 5 investigated the hypothesis that C37:4 alkenone concentrations can be used as a proxy for lake water salinity. We sampled a suite of lakes on the Chinese-

Tibetan plateau in the Lake Qinghai region, a region populated by lakes of various

salinities containing alkenone profiles with variable C37:4 concentrations. Our results

supported previous findings that C37:4 concentrations increased with decreasing salinity. However, instead of a single haptophyte species changing alkenone

production with varying salinity, we found evidence to suggest that variations in C37:4 alkenone concentrations are a reflection of changing haptophyte communities. We identified two haptophyte species in Lake Qinghai, each with a distinct alkenone signature. The fluctuation in abundance of these two haptophyte species explains the

observed variations in C37:4 concentrations. We conclude that environmentally- dictated variations in haptophyte communities are responsible for the observed

relationship between C37:4 alkenone concentrations with salinity.

In Chapter 6 we targeted the elusive C37:4 dominant alkenone-producing haptophyte from Lake George, ND. Previous research (Theroux et al., 2010; Toney et al., 2012) had shown that enrichment cultures inoculated with Lake George sediments

could reliably produce C37:4 alkenones. Relying upon a size fractionation experiment, we successfully isolated the C37:4-dominant haptophyte away from its C37:3-dominant

8 relative, and grew this novel C37:4-producing haptophyte in culture to determines its

k U37 calibration. The C37:4-dominant haptophyte was isolated in the <3µm fraction, and microscopy revealed this haptophyte is indeed a picoplankter. This finding improves our understanding of the ecology of this novel, and long-mysterious haptophyte species and provides insight into how its physiology gives it a competitive advantage in lake systems.

Chapter 7 was our most aggressive investigation into a haptophyte bloom event and alkenone production in a multiple-haptophyte system. With this study, we analyzed the geochemical and biological profiles in Lake George, ND throughout the course of the seasonal bloom cycle. Using biweekly sampling, we generated alkenone concentration profiles coupled with DNA profiling of eukaryotic microbial communities. This is the first study to couple high-throughput DNA sequencing with alkenone analyses to decipher changes in haptophyte community composition and biomarker production throughout a seasonal cycle. Additionally, we isolated individual haptophytes from Lake George and grew these novel species at controlled

k temperatures to define their U37 calibrations. Our multi-disciplinary approach revealed that in Lake George, the two-haptophyte system, at times, produces a composite alkenone signature recorded in the sediment record.

9 References

Brassell SC, Eglinton G, Marlowe IT, Pflaumann U, Sarnthein M. (1986). Molecular stratigraphy - a new tool for climatic assessment. Nature 320:129-133.

Brassell SC, Dumitrescu M. (2004). Recognition of alkenones in a lower Aptian porcellanite from the west-central Pacific. Organic Geochemistry 35:181-188.

Chu GQ, Sun Q, Li SQ, Zheng MP, Jia XX, Lu CF, Liu JQ, Liu TS. (2005). Long- chain alkenone distributions and temperature dependence in lacustrine surface sediments from China. Geochimica Et Cosmochimica Acta 69:4985-5003.

Conte MH, Thompson A, Lesley D, Harris RP. (1998). Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica. Geochimica Et Cosmochimica Acta 62:51-68.

Coolen MJL, Muyzer G, Rijpstra WIC, Schouten S, Volkman JK, Sinninghe Damsté J. (2004). Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth and Planetary Science Letters 223:225-239.

D'Andrea WJ, Huang Y. (2005). Long chain alkenones in Greenland lake sediments: Low δ13C values and exceptional abundance. Organic Geochemistry 36:1234- 1241.

D'Andrea WJ, Lage M, Martiny JBH, Laatsch AD, Amaral-Zettler LA, Sogin ML, Huang YS. (2006). Alkenone producers inferred from well-preserved 18S rDNA in Greenland lake sediments. Journal of Geophysical Research- Biogeosciences 111:doi:10.1029/2005JG000121.

Eltgroth ML, Watwood RL, Wolfe GV. (2005). Production and cellular localization of neutral long-chain lipids in the haptophyte algae Isochrysis galbana and Emiliania huxleyi. Journal of Phycology 41:1000-1009.

Liu H, Probert I, Uitz J, Claustre H, Aris-Brosou S, Frada M, Not F, de Vargas C. (2009). Extreme diversity in noncalcifying haptophytes explains a major pigment paradox in open oceans. Proceedings of the National Academy of Sciences of the United States of America 106:12803-12808.

Liu WG, Liu ZH, Fu MY, An ZH. (2008). Distribution of the C37 tetra-unsaturated alkenone in Lake Qinghai, China: A potential lake salinity indicator. Geochimica Et Cosmochimica Acta 72:988-997.

Marlowe IT, Green JC, Neal AC, Brassell SC, Eglinton G, Course PA. (1984). Long- Chain (n-C37-C39) alkenones in the prymnesiophyceae. Distribution of

10 alkenones and other lipids and their taxonomic significance. British Phycological Journal 19:203-216.

Pearson EJ, Juggins S, Farrimond P. (2008). Distribution and significance of long- chain alkenones as salinity and temperature indicators in Spanish saline lake sediments. Geochimica Et Cosmochimica Acta 72:4035-4046.

Prahl FG, Wolfe GV, Sparrow MA. (2003). Physiological impacts on alkenone paleothermometry. Paleoceanography 18:doi:10.1029/2002PA000803.

Prahl FG, Muehlhausen LA, Zahnle DL. (1988). Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochimica Et Cosmochimica Acta 52:2303-2310.

Prahl FG, Wakeham SG. (1987). Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330:367-369.

Rechka JA, Maxwell JR. (1988). Characterisation of alkenone temperature indicators in sediments and organisms. Organic Geochemistry 13:727-734.

Rontani JF, Beker B, Volkman JK. (2004). Long-chain alkenones and related compounds in the benthic haptophyte Chrysotila lamellosa Anand HAP 17. Phytochemistry 65:117-126.

Rontani JF, Prahl FG, Volkman JK. (2006). Re-examination of the double bond positions in alkenones and derivatives: biosynthetic implications. Journal of Phycology 42:800-813.

Sawada K, Handa N, Shiraiwa Y, Danbara A, Montani S. (1996). Long-chain alkenones and alkyl alkenoates in the coastal and pelagic sediments of the northwest North Pacific, with special reference to the reconstruction of Emiliania huxleyi and Gephyrocapsa oceanica ratios. Organic Geochemistry 24:751 - 764.

Schulz H-M, Schöner A, Emeis K-C. (2000). Long-chain alkenone patterns in the Baltic sea--an ocean-freshwater transition. Geochimica Et Cosmochimica Acta 64:469-477.

Sun Q, Chu GQ, Liu GX, Li S, Wang XH. (2007). Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellosa (Haptophyceae). Organic Geochemistry 38:1226-1234.

Theroux SM, D'Andrea WJ, Toney JL, Amaral-Zettler LA, Huang Y. (2010). Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes: Implications for continental paleotemperature reconstructions. Earth and Planetary Science Letters 300:311-320.

11

Thiel V, Jenisch A, Landmann G, Reimer A, Michaelis W. (1997). Unusual distributions of long-chain alkenones and tetrahymanol from the highly alkaline Lake Van, Turkey. Geochimica Et Cosmochimica Acta 61:2053-2064.

Toney JL, Huang Y, Fritz SC, Baker PA, Grimm E, Nyren P. (2010). Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes in the interior United States. Geochimica Et Cosmochimica Acta

k' Versteegh GJM, Riegman R, de Leeuw JW, Jansen JHF. (2001). U37 values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations. Organic Geochemistry 32:785-794.

Zink KG, Leythaeuser D, Melkonian M, Schwark L. (2001). Temperature dependency of long-chain alkenone distributions in Recent to fossil limnic sediments and in lake waters. Geochimica Et Cosmochimica Acta 65:253-265.

12 CHAPTER 2

PHYLOGENETIC DIVERSITY AND EVOLUTIONARY RELATEDNESS OF

ALKENONE PRODUCING HAPTOPHYTE ALGAE IN LAKES: IMPLICATIONS

FOR CONTINENTAL PALEOTEMPERATURE RECONSTRUCTIONS

SUSANNA THEROUX1,2

WILLIAM D’ANDREA1,*

JAIME TONEY1,‡

LINDA AMARAL-ZETTLER1,2,3

YONGSONG HUANG1

1 BROWN UNIVERSITY, Department of Geological Sciences

2 MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

3 BROWN UNIVERSITY, Department of Ecology and Evolutionary Biology

* Present address: LAMONT-DOHERTY EARTH OBSERVATORY, Biology and

Paleo Environment

‡ Present address: UNIVERSITY OF GLASGOW, School of Geographical and Earth Sciences

Published in: Earth and Planetary Science Letters 300, 311-320.

13 Abstract

Alkenones have been found in an increasing number of lakes around the world, making them a promising new tool for continental paleoclimate reconstruction. However, individual lakes may harbor different species of haptophyte algae with different sensitivities to temperature variations, thus presenting a significant challenge to the use of lacustrine alkenones for paleotemperature reconstructions. To explore the extent of lacustrine haptophyte diversity, we conducted the first comprehensive phylogenetic and geochemical survey of lacustrine alkenone producers. We sampled

15 alkenone-containing lake surface sediments from a variety of geographic locales and inferred identities of environmental sequences using 18S ribosomal RNA (rRNA) gene based phylogenies. For two lakes, BrayaSø in southwest Greenland and Tso Ur on the Tibetan Plateau, we also analyzed both surface and down-core sediments to characterize haptophyte populations through time. In parallel with phylogenetic analyses, we determined the alkenone distributions (including C37/C38 ratios, and the presence/absence of C38 methyl ketones and C40 compounds) in all the samples. The resulting alkenone profiles from this study do not all align with traditional “marine” versus “coastal/lacustrine” alkenone profiles. Additionally, our genetic data indicate the presence of multiple haptophyte species from a single lake sediment sample; these composite haptophyte populations could not be discerned from the alkenone profiles alone. These results show that alkenone profiles are not a reliable way to assess the haptophyte algae in lakes and that DNA fingerprinting is a preferred approach for species identification. Although closely-related haptophyte species or subspecies may not warrant different temperature calibrations, our results emphasize the importance of genetic data for inferring haptophyte identities and eventually selecting alkenone- temperature calibrations for paleoclimate reconstructions.

14 1. Introduction

Marine haptophyte algae produce unsaturated long-chain alkenones (LCAs)

K K' that serve as a well-established paleotemperature proxy (U37 or U37 ) in marine sediments (Volkman et al., 1980; Marlowe et al., 1984; Brassell et al., 1986; Prahl and Wakeham, 1987; Conte et al., 2006). Alkenones! are! produced by a limited number of haptophyte algal species within the order Isochrysidales (Müller et al.,

1998; Medlin et al., 2008), notably Emiliania huxleyi and Gephyrocapsa oceanica in the open ocean (Volkman et al., 1980; Marlowe et al., 1984), and Isochrysis galbana

(Marlowe et al., 1984; Versteegh et al., 2001) and Chrysotila lamellosa in coastal regions (Volkman et al., 1980; Marlowe et al., 1984; Volkman et al., 1995; Conte et al., 1998; Rontani et al., 2004). Emiliania huxleyi and G. oceanica (Müller et al.,

1998; Conte et al., 2006) are the dominant alkenone-producers in marine surface

K' waters, enabling the application of a global U37 calibration to marine sea surface temperature (SST) reconstructions.

Long chain alkenones (LCAs)! in lake sediments have great potential for continental climate reconstructions. LCAs have been increasingly reported in lakes around the world, including Antarctica, Europe, Greenland, South America, China and the western United States (Zink et al., 2001; Coolen et al., 2004; D'Andrea and

Huang, 2005; Liu et al., 2008; Pearson et al., 2008; Toney et al., 2009). Lacustrine alkenone unsaturation ratios correspond to lake seasonal surface water temperatures and mean annual air temperature (Zink et al., 2001; Chu et al., 2005; D'Andrea and

Huang, 2005; Sun et al., 2007). However, individual alkenone-containing lakes may require different calibrations, depending on the haptophyte species present (Prahl and

15 Wakeham, 1987; Prahl et al., 1988; Volkman et al., 1995; Versteegh et al., 2001;

Zink et al., 2001; Rontani et al., 2004; Chu et al., 2005; Sun et al., 2007; D'Andrea,

2008; Liu et al., 2008). Previous studies have generated a variety of LCA profiles from different lakes (Liu et al., 2008; Pearson et al., 2008), with varying ratios of

C37/C38 alkenone, percent concentrations of tetraunsaturated C37 ketones (% C37:4), and the presence/absence of C38 methyl ketones and C40 alkenones. These variable distributions suggest that alkenones from different lakes could be produced by a diversity of haptophyte species, and therefore a single global temperature calibration may not be viable.

Although alkenone signatures provide a simple way to compare alkenones from different lakes, they do not provide definitive taxonomic information at the species level. Recent studies identified haptophyte species using genetic methods and enabled the identification of novel haptophytes species without the need for cultivation. Work by Coolen et al. (2004) targeted the 18S rRNA molecule to identify the haptophyte responsible for alkenone production in ancient Ace Lake, Antarctic sediments. These authors identified preserved DNA of a haptophyte closely related to

I. galbana. D’Andrea and Huang (2005) reported the occurrence of a novel bloom- forming haptophyte species in lakes of Southwestern Greenland with extremely high alkenone concentrations, characterized by an abundance of tetra-unsaturated

alkenones and the presence of C38 methyl and ethyl ketones. Genetic work determined this Greenland haptophyte forms a distinct phylogenetic group from the marine and lacustrine phylotypes within the Isochrysidales (D'Andrea et al., 2006).

16 This paper aims to expand the phylogenetic dataset of lacustrine alkenone- producing haptophytes and compare 18S rRNA gene-based phylogenetic relationships with the observed alkenone lipid signatures. We studied lake sediments from Greenland, China, Northern Canada, and the continental United States. We hypothesized that different alkenone lipid signatures in lake sediments can be attributed to the genetic diversity of the haptophyte species, and that multiple alkenone-producing haptophyte species could be present in individual lakes. With this comprehensive survey we were able to identify a) presence/absence of biogeographical trends in lacustrine haptophyte distribution, b) relationships between alkenone lipid profiles and genetic rRNA-inferred phylogenies, and c) the accuracy and fidelity of the alkenone signature as a haptophyte identification tool. This dataset of haptophyte alkenone signatures and corresponding species classifications assesses the reliability of haptophyte species identification through alkenone signature alone.

2. Methods

2.1 Sampling

Previous studies have demonstrated that DNA of haptophyte algae can be readily extracted from lake sediments (Coolen et al., 2004; D'Andrea et al., 2006).

We analyzed sediments from 15 lakes (Figure 1, Table 1) from the continental

United States, Canada, Greenland, and Tibet. Lake surface sediment samples were collected by gravity core, split, and the upper 1 cm was used for analysis. Lakes

Clear, Medicine, George, and Skoal sediments were freeze-dried before DNA

17 extraction. All other sediment samples were kept at 4 °C and in the dark until processing. Individual sediment samples were split for DNA and alkenone analysis.

For Lake BrayaSø water column DNA, we collected water on June 26, 2007 at 10 m depth, corresponding to the chlorophyll maximum that day. One liter was filtered using a 0.22 µm pore size Sterivex filter (Millipore, Bedford, MA, USA), treated with Puregene cell lysis buffer (Qiagen, Carlsbad, CA, USA) and kept cool and in the dark until freezing at -20 °C.

2.2 DNA extraction

We extracted DNA from 1 g of lake surface sediment for each lake we examined using MoBio UltraPure DNA Extraction Kit for sediments (Qiagen,

Carlsbad, CA, USA) according to the manufacturer’s instructions. For lakes BrayaSø and Tso Ur, down-core sediment samples were also extracted for DNA analysis. For water samples, Sterivex filters were extracted using a Puregene Cell Kit (Qiagen) according to the manufacturers instructions. Genomic DNA was polyethylene glycol

(PEG) purified to remove proteins and other contaminants that inhibit PCR reactions:

DNA was suspended in PEG at 4 °C overnight, centrifuged, and the pellet rinsed with ethanol (LaMontagne et al., 2002). We quantified total extracted genomic DNA yields using a NanoDrop nucleic acid spectrophotometer (Thermo Scientific,

Wilmington, DE).

2.3 DNA amplification and sequencing

18 We amplified genomic DNA using haptophyte-specific oligonucleotide

(Simon et al., 2000; Coolen et al., 2004) primers targeting 18S rRNA coding regions.

Forward and reverse primers corresponded to E. coli 18S rRNA positions 429 and

887, respectively (Table 2). Polymerase chain reactions (PCRs) were performed on an Eppendorf Gradient Thermocycler (Hamburg, Germany) with the following conditions after D’Andrea et al. (2006): 4 min initial denaturing at 96 °C, 35 cycles of denaturing for 30 s at 94 °C, followed by 40 s primer annealing at 55 °C and primer extension 40 s at 72 °C, with a final extension of 10 min at 72 °C. We devoted a separate PCR run for each lake to avoid cross contamination. We ran each sample at optimal dilution (1:10, 1:100, or 1:1000) for maximum product yield. PCR reactions were run in 50 µl volume using Promega GoTaq polymerase (Madison, WI, USA).

Triplicate PCR products were pooled for each sample, and templates were purified using an Invitrogen PureLink Purification kit (Carlsbad, CA, USA) with the high-cut- off binding buffer to eliminate fragments < 200 bp. Purified PCR products were A- tailed and purified using the PureLink high cut-off kit. Cloning was performed using the Invitrogen TOP10 cloning kit with electrocompetent cells. Protocol was according to the manufacturer’s instructions. One hundred clones were picked for each sample.

Plasmid DNA was isolated using a RevPrep Orbit robotic template preparation instrument (Genomic Solutions, Ann Arbor, MI), and prepared templates were sequenced on an ABI 3730XL (Applied Biosystems, Foster City, CA) capillary sequencer using the BigDye protocol with universal M13 forward and reverse primers according to the manufacturer’s instructions. All sequencing was performed at the

19 Marine Biological Laboratory W. M. Keck Ecological and Evolutionary Genetics

Facility.

For additional PCR amplification of the complete Greenland Lake BrayaSø haptophyte 18S rRNA gene, we designed an internal primer and paired this with universal eukaryotic forward and reverse primers (Medlin et al., 1988), in addition to the haptophyte-specific primers (Table 2). The PCR program was as follows: 2 min denaturing at 94 °C, 30 cycles of denaturing for 1 min at 94 °C, primer annealing at

45 °C for 1 min, and primer extension for 2 min at 72 °C. The final extension was for

10 min at 72 °C. DNA purification, cloning, and sequencing were as above.

2.4 Bioinformatics and phylogenetic reconstructions

A bioinformatics pipeline using the programs phred, cross-match, and phrap, translated chromatograms into base-calls and associated quality scores, removed vector sequences and assembled forward and reverse reads into full-length sequences for each of the cloned PCR amplicons (Ewing and Green, 1998; Ewing et al., 1998;

Lasek-Nesselquist et al., 2008). Only sequences greater than 400 bp and with a complete forward and reverse primer were retained. Base-calls were verified and sequences were manually edited with the program Consed (Gordon et al., 1998) for chromatogram viewing. We screened edited sequences for chimeras using the computer program Mallard (Ashelford et al., 2006). The number of sequences recovered for each sample is listed in Table 3. For samples with only single read coverage, we verified the sequence by re-sequencing the template in both the forward and reverse directions. We identified closest relatives within the GenBank database

20 using BLAST (Altschul et al., 1997). Assembled sequences were aligned using the

ARB software program v. 07.07.11 (Ludwig et al., 2004) against the April 2008 Silva

94 Ref database (Pruesse et al., 2007) using the FastAligner option followed by manual adjustment. Operational Taxonomic Units (OTUs) were obtained using the

DOTUR program (Schloss and Handelsman, 2005) furthest neighbor algorithm setting at 95 % cut-off criterion. Representative sequences for each OTU were selected for tree construction. Full-length 18S rRNA gene contigs were assembled using the MacClade software package version 4 (Maddison and Maddison, 2000). We constructed a distance matrix by calculating differences/sequence length for aligned reference haptophyte sequences and representative OTU sequences, using the neighbor joining method as implemented in ARB.

A custom filter of 1,756 positions was constructed manually for both Bayesian analyses of 18S rRNA sequences. We selected Cyclonexis annularis, Chrysoxys sp.,

Ochromonas danica, Odontella sinensis and Thraustochytrium multirudimentale as outgroups for our Bayesian analyses after de Vargas et al. (2007). We subjected our datasets to a Bayesian analysis using MrBayes version 3.0b4 (Ronquist and

Huelsenbeck, 2003) under the GTR model of substitution (Lanave et al., 1984;

Tavaré, 1986; Rodriguez et al., 1990) considering invariants and a gamma-shaped distribution of the rates of distribution among sites. The chain length for our analysis was 1,000,000 generations with trees sampled every 100 generations using MCMC

(Markov Chain Monte Carlo) analysis. The first 10,000 trees were discarded as burn- in for the tree topology and posterior probability. OTU representative sequences and full-length 18S rRNA gene sequences from reference taxa were analyzed to infer

21 OTU species’ identities, and included a total of 92 taxa in our analysis. We further analyzed the full-length 18S rRNA molecule for the BrayaSø Greenland haptophyte and aligned this against a full-length haptophyte reference taxa dataset consisting of

220 taxa.

2.5 Lipid extraction and analysis

Alkenone extraction was after D’Andrea and Huang (2005). Alkenone samples came from the same sample used for DNA work. We freeze-dried and homogenized samples manually. We extracted samples with 9:1 Dichloromethane

(DCM) : Methanol (MeOH) using an Accelerated Solvent Extractor ASE200

(Dionex, Sunnyvale, CA, USA). Extracts were separated into acid and neutral fractions using a solution of DCM:Isopropyl alcohol 2:1 (v/v). The neutral fraction was further separated into aliphatic (hexane elution), ketone (DCM), and alcohol

(ethyl acetate: hexane 1:3) fractions using a flash silica gel column. The DCM fraction was analyzed using an Agilent 6890plus Gas Chromatograph Flame Ionization

Detector (GC-FID) for quantification. Organic compounds were verified by running on the GC-MS (Agilent 6890N Gas Chromatograph coupled to an Agilent 5973

Network Mass Selective Detector) with the same temperature program and chromatograms were compared to mass spectral data from previously reported standards and GC retention times (de Leeuw et al., 1980; Marlowe et al., 1984). LCA concentrations were determined from GC-FID analysis of the ketone fractions based

on an internal C36 alkane standard.

22 3. Results

3.1 Phylogenetic relationships among alkenone-producing haptophytes

Ten out of fifteen lake sediments yielded positive amplification for haptophytes. Five lakes (Lake Qinghai, Brush Lake, Moon Lake, Mahoney Lake, and

GaHai) yielded genomic DNA (DNA extracts >5 ng/µl; data not shown) but were negative for amplification using haptophyte-specific primers. Except Lake Qinghai, with a previously published alkenone profile (Li et al., 1996), the lakes with no haptophyte DNA amplification also had below detection or convoluted alkenone profiles, indicative of low haptophyte cell numbers in the surface sediments. From the ten lake sediments with positive haptophyte DNA amplification, we identified eight

OTUs at 95 % similarity from the partial 18S rRNA gene data. Individual lakes had multiple OTUs present, and the percent distribution for each lake is outlined in Table

3a.

We constructed a phylogeny using a representative sequence for each OTU along with previously published haptophyte 18S rRNA genes (Fig. 2). A posterior probability value of 1.0 supports the monophyly of the Isochrysidales order, in which all OTU sequences are grouped. The alkenone-producing haptophyte sequences clustered into three groups. Haptophyte Group I was comprised of representative sequences from OTUs 1, 2, 3, 4 and 5, and previously published sequences from the

Greenland lakes (D'Andrea et al., 2006), with strong posterior probability support

(1.0). We also identified these Group I OTUs in lake sediments from BrayaSø (BS),

Skoal (SK), Upper Murray (UM), Tso Ur downcore (TUd), Keluke (KE), and

Pyramid (PY) lakes. Group II was comprised of OTU 6, OTU 7 and OTU 8

23 sequences, in addition to C. lamellosa, I. litoralis, Dircrateria sp. sequences, and sequences from Ace Lake, Antarctica (Coolen et al., 2004). Sequences OTU 6, OTU

7 and OTU 8 appeared in Medicine (MD), George (LG), Skoal, Great Salt (GS),

Pyramid, Keluke Hu and Tso Ur surface sediments (TUs). These sequences formed a sister group to marine species E. huxleyi, G. oceanica, and an unidentified marine coccoid haptophyte (U40924) comprise Group III.

A pairwise comparison among the eight OTU representative sequences and the Greenland haptophyte full-length sequence showed that OTU 1 and the Greenland

Lake BrayaSø haptophyte sequence, as well as OTU 5 and the BrayaSø haptophyte sequence, were the most similar, with only 0.2 % differences (Table 4a). OTU 2 sequences were found in sediments from BrayaSø as well as Tibetan plateau Tso Ur, and OTU 5 sequences were found in BrayaSø, as well as Keluke Hu, Pyramid lake, and Upper Murray Lake. OTU 2 and OTU 6 representative sequences had the greatest distance, with 10.2 % difference. Notably, OTU 6 was found only in Tso Ur surface sediments and OTU 2 was found in Tso Ur ancient sediments.

According to the full-length 18S rRNA gene-based phylogenies (Fig. 3), the

Greenland haptophyte (extracted from the BrayaSø water column) branched basal to a lineage of coastal/lacustrine haptophytes I. galbana and C. lamellosa, and shared common ancestry with marine species E. huxleyi and G. oceanica (posterior probability of 0.92). Table 4b shows the pairwise distances among full-length 18S rRNA sequences for the reference taxa and the BrayaSø water haptophyte. The

BrayaSø haptophyte sequence shared the percent similarity with the E. huxleyi

(AF184167) and G. oceanica (AJ246276) sequences, with only 1.9 % difference,

24 while it was the most divergent from C. lamellosa (AM490998) with 4.1 % difference.

3.2 Alkenone signatures of individual lake sediments

In order to numerically compare the alkenone distributions in different

samples, we computed the relative abundances of C37 to C38 alkenones (Prahl et al.,

1988; Conte et al., 1998; Chu et al., 2005), the presence/absence of C38 methyl ketones (Conte et al., 1994; Schulz et al., 2000; Bendle et al., 2005), the

presence/absence of C40 compounds, and the percent concentration of C37:4 relative to total C37 alkenones. Three exemplary alkenone signatures of individual lake

sediments are presented in Figure 4, and corresponding C37/C38 values are given in

Table 5b. Figure 4a shows the alkenone gas chromatogram for Upper Murray lake

surface sediments. These alkenones had a C37/C38 ratio of 2.4, the presence of the C38 methyl ketone, and trace amounts of the C40 compound (Table 5b). The presence of

the C38 methyl ketone was similar to marine alkenone signatures from E. huxleyi and

BrayaSø alkenone signatures (D'Andrea and Huang, 2005). Also like BrayaSø, the

Upper Murray sediments had a high percentage of C37:4 (63 %) and both lake had haptophyte sequences that branched with Group I (Table 5). Figure 4b shows the

alkenone profile for Medicine Lake with a C37/C38 ratio of 3.8, very low concentrations of C38 methyl ketones, and the presence of the C40 compound (Table

5b). The predominance of the ethyl ketone and the high C37/C38 ratio was similar to I. galbana alkenone signatures, and accordingly, Medicine Lake haptophyte sequences branched with I. galbana sequences within Group II (Table 5a). Figure 4c depicts

25 the alkenone profile for Pyramid lake surface sediments, with a C37/C38 ratio of 1.9, the presence of the methyl ketone and trace levels of C40 compounds (Table 5b).

Pyramid lake sediments had a predominance of C37:3 alkenones, and a percent C37:4 of

22%, and DNA sequences branched with both Group I and Group II of the

Isochrysidales.

The majority of the lake sediments analyzed in this study had a predominance

of the tetraunsaturated C37 alkenone (Table 5b). Elevated percent concentrations of

C37:4 have been proposed as characteristics for lacustrine Chrysotila and Isochrysis species (Conte et al., 1994), and have been suggested as a paleosalinity proxy, with

low salinity corresponding to high percent C37:4 (Liu et al., 2008). Sediments from

Great Salt Lake, Pyramid Lake, and Tso Ur had higher concentration of tri-

unsaturated ketones. The lake sediments with the highest percentage of C37:4 alkenones was Medicine lake, with an average salinity of 9, and the lake with the

lowest percentage of C37:4 alkenones was Pyramid Lake, NV, with a salinity of 5.

Given that the salinity for the lakes in this study ranged from 0.04 to 270, our results

therefore do not support a linear relationship between salinity and percent C37:4 across multiple lakes and species.

3.3 Downcore alkenone and OTU distribution in BrayaSø and Tso Ur

Surface sediment alkenone signatures from both lakes BrayaSø and Tso Ur highly resembled ancient down-core sediment alkenone signatures from the lakes.

BrayaSø alkenone signatures were characterized by a high tetra-unsaturated (C37:4) alkenone concentration, the presence of the C38 methyl ketone, a C37/C38 ratio of < 2,

26 and the absence of the C40 compounds in both surface and downcore sediments

(D'Andrea and Huang, 2005). Table 3b shows the down-core variation in BrayaSø

OTU sequence distribution. Spanning 20 cm to 100 cm depth in the core, or roughly

1,370 to 6,200 years BP, there was little change in the distribution of OTUs. As shown in Table 3a, the maximum percent dissimilarity among these OTU representative sequences was 6.5 % (OTU 4 and OTU 2), the minimum difference was 0.4 % (OTU 1 and OTU 5).

Tso Ur sediments were extracted at 0 cm and 20 cm core depth (age

unknown). Tso Ur surface and ancient sediments were very similar with C37/C38 values of < 2, the presence of C38 methyl ketones, and the presence of C40 compounds

(Table 5b). Surprisingly, the OTU sequence distribution was very dissimilar, with sequences from Tso Ur surface sediments falling into Group II, and sequences from

Tso Ur downcore sediments falling into Group I (Table 3a).

4. Discussion

4.1 OTU distribution and alkenone signatures

All haptophyte sequences in this study grouped within the order

Isochrysidales. Previous tenet predicted distinct alkenone signatures according to marine or lacustrine species representation (Conte et al., 1998; Prahl et al., 1988; Chu et al., 2005; Liu et al., 2008). However, our results did not corroborate this pattern.

Our results confirm our hypothesis that alkenone signatures alone do not adequately reflect the extent of haptophytes species diversity in various lakes: genetic data provided a more comprehensive measurement of the haptophyte species community.

27 Lakes Clear, Great Salt, George and Medicine contained haptophyte sequences that branched with haptophyte Chrysotila and Isochrysis species. Upper Murray lake also contained Group I representation. Pyramid and Skoal lakes yielded haptophyte sequences that branched into both Groups I and II, and TsoUr yielded Group I sequences in the ancient sediments and Group II sequences in the recent sediments.

We evaluated C37/ C38 ratios and determined there was no clear pattern with phylogenetic placement. While lacustrine profiles often displayed a higher C37/ C38 ratio than marine profiles, Conte et al. (1998) found an increase in C38 ethyl

alkenones in comparison to C38 methyl alkenones during late logarithmic and stationary growth phase of E. huxleyi. We therefore conclude that C37/ C38 ratios do not give an accurate indication of haptophyte species identity, and may reflect algal

growth status at alkenone biosynthesis. The presence of the C38 methyl ketone was previously attributed to marine haptophyte presence (Schulz et al., 2000); with this

dataset we show the presence of the C38 methyl ketone in multiple lakes with Group II haptophytes (e.g. Great Salt lake, Pyramid Lake), i.e. haptophytes more closely related to lacustrine Isochrysis and Chrysotila species. In all samples, the presence of

the C38 methyl ketone was confirmed with mass spectral analysis, although the relative abundance of C38 methyl versus ethyl varied (data not shown). The C40 alkenone was found in trace amounts in four of our lakes. With the exception of

ancient TsoUr sediments, the C40 compound was only found in sediments with haptophyte species branching among Group II. While not all Group II haptophytes

appeared to produce C40 compounds, the presence of C40 may indicate the presence of a Group II haptophyte species. In the downcore sediments of TsoUr, the Group II

28 haptophytes (possible source for C40 alkenones) may be less abundant and were therefore undetected from older sediments.

One of the significant observations from this study was the occurrence of

DNA from multiple haptophyte species in a single lake. Lakes Skoal and Pyramid, and Keluke Hu ancient sediments highlight the importance of verifying haptophyte species identities with genetic data, as opposed to alkenone signature alone. For these three sites, the alkenone signature may have been representative of a single haptophyte species, or the combined signature from two haptophyte species. It is also possible one or more of the haptophyte species does not contribute to the alkenone profile. Verifying the contributors to this alkenone profile is of great importance

K' when selecting an appropriate U37 calibration for reconstructing water surface temperature. Likewise, both TsoUr recent and ancient sediments displayed very similar alkenone signatures.! Upon DNA sequencing, it was revealed that sequences from two separated haptophyte groups were amplified from the each sample. These results could arise if 1) the haptophyte that dominates the alkenone signature (by producing large quantities of alkenones) is not the same as the haptophyte dominating

DNA production or 2) environmental conditions, rather than phylogeny, determine alkenone lipid signatures. In contrast, Lake BrayaSø had both conserved alkenone signatures and haptophyte species representation over the past 6 ka, deeming this lake

K an excellent candidate for U37 based temperature reconstructions.

4.2. Haptophyte !diversity and genetic markers

29 Sequence recovery varied among samples, and may be due to poor DNA preservation conditions in situ, or a poor cloning success rate. Our sequence-yield numbers therefore do not represent haptophyte DNA concentration, but number of successful sequences recovered. According to our partial 18S rRNA gene sequencing, no lake with > 5 sequences recovered contained a single dominant haptophyte species. The presence of multiple OTUs in single lakes suggested a significant degree of microheterogeneity in the lake haptophyte populations. Our haptophyte-specific primers targeted the most variable region of the haptophyte 18S rRNA gene.

Incorporating a larger region of the 18S rRNA molecule enabled an improved support for the position of the novel Greenland Lake BrayaSø haptophyte species within the

Isochrysidales (posterior probably increased from 0.66 to 0.92; Figure 3). Additional sequencing of the large subunit rRNA gene, plastid genes, or COI genes (Coolen et al., 2009) may further clarify poorly resolved haptophyte phylogenies.

4.3 Geographical distribution and environmental controls

The OTU distributions did not display a strong geographical signal. Most

OTUs were represented in multiple lakes, e.g. OTUs 7 and 8 were present in more than half of the lakes analyzed. OTU 3 and OTU 4 only occurred in Lake BrayaSø and OTU 6 only occurred in TsoUr. Given the ubiquitous distribution of these haptophyte OTUs, haptophyte species’ dispersal mechanisms must be able to achieve great distance. Whether by biotic or abiotic (winds, clouds) mechanisms, these haptophyte species may potentially encyst to survive transport. The haptophyte species Prymnesium parvum is known to form cysts (Pienaar, 1980), and given the

30 absence of haptophyte cells in the water column during the non-bloom periods in

Greenland, these cells likely encyst or enter a vegetative stage in their life cycle.

Marine aerosols can also transports organic compounds into the lake (Sicre et al.,

1990), but this would not explain the presence of haptophyte DNA, and therefore we assume the alkenones present in the lake are synthesized in situ.

The Greenland phylotype shared a common ancestor with the

Isochrysis/Chrysotila group, which together formed a sister group with the marine haptophytes E. huxleyi and G. oceanica. The BrayaSø alkenone signatures were

similar to traditional marine alkenone signatures (i.e. presence of C38Me, absence of

C40, low C37/C38), with the exception that BrayaSø contained high tetraunsaturated alkenone concentrations. The conditions of the ancestral Lake BrayaSø species are unclear, as the lake was not formed by marine capture (Anderson and Bennike, 1997;

D'Andrea and Huang, 2005) and there is no evidence for alkenone lipids in the first

2000 years of lake sediments. The full-length 18S rRNA sequence phylogeny showed that the BrayaSø haptophyte diverged before the lacustrine and coastal I. galbana and

C. lamellosa species. According to this phylogeny, the last common ancestor to the

BrayaSø haptophyte and I. galbana and C. lamellosa may have been marine. If we attribute the high tetraunsaturated alkenones in lake BrayaSø sediments to very cold temperature waters and/or low salinity, the BrayaSø alkenone signature has been conserved from its marine ancestor.

Lake BrayaSø formed approximately 8,000 years before present (BP)

(McGowan et al., 2003), and the occurrence of alkenones in the sediments began ~

2,000 years later at a time of increased lake water salinity (McGowan et al., 2003;

31 Anderson and Leng, 2004). Salinity has previously been suggested as a pre-requisite for alkenone-producing haptophyte populations (Rosell-Melé, 1998; Schulz et al.,

2000). Nine of ten lakes in our study were either oligosaline, brackish or alkaline and, notably, Upper Murray is a freshwater lake (Table 1). It is possible that the alkenone- producing haptophytes require a saline environment for survival, or perhaps the haptophyte communities are present but only produce alkenones when the water reaches a certain salinity threshold. Salinity gradient culture experiments using novel haptophyte species would help clarify this assertion.

5. Conclusions

Haptophyte communities in lakes can be precisely determined using genetic fingerprints preserved in lake sediments and collected from the water column during algal growing seasons. Alkenone profiles, however, are not a reliable way to determine the haptophyte species in lakes: similar alkenone profiles may be produced by different species, whereas different alkenone profiles may be derived from similar haptophyte species. We identified eight operational taxonomic units (OTUs) from a series of ten lakes, and generated a haptophyte phylogeny that further resolved the evolutionary relationships of alkenone-producing haptophytes. Although different

OTUs do not necessarily imply differential alkenone-temperature sensitivity, the correct identification of haptophyte community members is critical for ultimately

K selecting the appropriate U37 calibration for both sea and lake surface temperature reconstructions. This study examined surface sediments in an order to capture the most-recent members! of the haptophyte community; however, if there were any

32 degradation between the overlying water column and the sediments, these haptophyte members would not be identified. The best alkenone-temperature calibration for individual lacustrine haptophyte species will eventually require isolation and culturing of algae under laboratory-controlled conditions.

Our lacustrine haptophyte sequences branched within the Isochrysidales, as expected for alkenone-producers, and none of our representative sequences branched directly with marine species E. huxleyi and G. oceanica. However, traditional lacustrine vs. marine alkenone profile trends were not conserved in our lacustrine

dataset: C37/C38 ratios ranged from 0.9 to 3.8, therefore spanning both the expected marine and lake ratios, and C38 methyl ketones were almost ubiquitously present, previously associated with marine profiles. The most prominent feature of lacustrine alkenones relative to marine alkenones is the predominance of the tetraunsaturated

C37 alkenones, which may have resulted from low salinity in lakes. Further research, preferably using algal cultures, will be needed to determine if low salinity conditions influence alkenone biosynthesis.

Sequencing the full-length 18S rRNA molecule sequence improved the resolution of the phylogenetic placement of the novel Greenland BrayaSø haptophyte species: the BrayaSø haptophyte shared a common ancestry with marine species E. huxleyi and G. oceanica, and branched basal to coastal/lacustrine species C. lamellosa and I. galbana. This result is not surprising given the similarity in alkenone signatures between the Greenland species and marine haptophytes, both with the presence of a

C38 methyl ketone and the absence of a C40 ketone. The Greenland down-core samples have a well-conserved alkenone signature through time, although we detected five

33 separate OTUs in these samples. All OTUs branch into Group I, suggesting the species are closely-related, if not subspecies or single species with microheterogeneity in their copies of 18S rRNA genes. This result suggests that when

K species are closely-related, a single U37 calibration is adequate for paleoclimate reconstructions, which can be achieved by performing an in situ calibration of the alkenone unsaturation index! (D'Andrea, 2008; Toney et al., 2009), or culture growth under controlled conditions. A continued appraisal of haptophyte species distributions, and the environmental constraints on alkenone production in a global diversity of lakes will greatly enhance our understanding of haptophyte community success and environmental controls on alkenone biosynthesis.

Acknowledgements

We would like to thank Marcelo Alexandre, Rafael Tarozo, and the members of the

W. M. Keck Ecological and Evolutionary Genetics Facility for lab support. We would also like to thank Ian Walker for contributing Mahoney Lake samples, and

Xiangdong Yang for the Tibetan plateau samples. We are grateful for financial support from the US National Science Foundation grant OPP 0520718 to Y. Huang, and for graduate student support to S. Theroux provided by the Brown University

Strominger Fellowship.

34 References

Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J. H., Zhang, Z., Miller, W., Lipman, D. J., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389-3402.

Anderson, N. J., Bennike, O., 1997. Holocene lake sediments in West Greenland and their palaeoclimatic and palaeoecological implications Geology of Greenland Survey Bulletin 176, 89-94.

Anderson, N. J., Leng, M. J., 2004. Increased aridity during the early Holocene in West Greenland inferred from stable isotopes in laminated-lake sediments. Quaternary Science Reviews 23, 841-849.

Ashelford, K. E., Chuzhanova, N. A., Fry, J. C., Jones, A. J., Weightman, A. J., 2006. New screening software shows that most recent large 16S rRNA gene clone libraries contain chimeras. Appl. Environ. Microbiol. 72, 5734-5741.

Bendle, J., Rosell-Mele, A., Ziveri, P., 2005. Variability of unusual distributions of alkenones in the surface waters of the Nordic seas. Paleoceanography 20. doi 10.1029/2004pa001025.

Besonen, M. R., Patridge, W., Bradley, R. S., Francus, P., Stoner, J. S., Abbott, M. B., 2008. A record of climate over the last millennium based on varved lake sediments from the Canadian High Arctic. Holocene 18, 169-180.

Brassell, S. C., Eglinton, G., Marlowe, I. T., Pflaumann, U., Sarnthein, M., 1986. Molecular stratigraphy - a new tool for climatic assessment. Nature 320, 129- 133.

Chu, G. Q., Sun, Q., Li, S. Q., Zheng, M. P., Jia, X. X., Lu, C. F., Liu, J. Q., Liu, T. S., 2005. Long-chain alkenone distributions and temperature dependence in lacustrine surface sediments from China. Geochimica Et Cosmochimica Acta 69, 4985-5003.

Conte, M. H., Sicre, M. A., Ruhlemann, C., Weber, J. C., Schulte, S., Schulz-Bull, D., Blanz, T., 2006. Global temperature calibration of the alkenone unsaturation k' index U37 in surface waters and comparison with surface sediments. Geochemistry Geophysics Geosystems 7. doi:10.1029/2005gc001054.

Conte, M. H., Thompson, A., Lesley, D., Harris, R. P., 1998. Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica. Geochimica Et Cosmochimica Acta 62, 51-68.

35 Conte, M. H., Volkman, J. K., Eglinton, G. 1994. Lipid biomarkers of the Haptophyta. In: Green, J. C. & Leadbetter, B. S. C. (eds.) The Haptophyte Algae. Clarendon Press. 351-377.

Coolen, M. J. L., Muyzer, G., Rijpstra, W. I. C., Schouten, S., Volkman, J. K., Sinninghe Damsté J., 2004. Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth and Planetary Science Letters 223, 225-239.

Coolen, M. J. L., Saenz, J. P., Giosan, L., Trowbridge, N. Y., Dimitrov, P., Dimitrov, D., Eglinton, T. I., 2009. DNA and lipid molecular stratigraphic records of haptophyte succession in the Black Sea during the Holocene. Earth and Planetary Science Letters 284, 610-621.

D'Andrea, W. J. 2008. Development and application of lacustrine alkenone paleothermometry in southwestern Greenland. PhD dissertation. Brown University, Providence. 197.

D'Andrea, W. J., Huang, Y., 2005. Long chain alkenones in Greenland lake sediments: Low δ13C values and exceptional abundance. Organic Geochemistry 36, 1234-1241.

D'Andrea, W. J., Lage, M., Martiny, J. B. H., Laatsch, A. D., Amaral-Zettler, L. A., Sogin, M. L., Huang, Y. S., 2006. Alkenone producers inferred from well- preserved 18S rDNA in Greenland lake sediments. Journal of Geophysical Research-Biogeosciences 111. doi: 10.1029/2005JG000121. de Leeuw, J. W., van der Meer, F. W., Rijpstra, W. I. C., Schenck, P. A., 1980. On the occurrence and structural identification of long chain unsaturated ketones and hydrocarbons in sediments. Physics and Chemistry of the Earth 12, 211- 217. de Vargas, C., Aubry, M. P., Probert, I., Young, J. 2007. Origin and evolution of coccolithophores: From coastal hunters to oceanic farmers. In: Falkowski, P. & Knoll, A. H. (eds.) Evolution of Aquatic Photoautotrophs. Elsevier Academic, New York. 251-281.

Donovan, J. J., 1994. On the measurement of reactive mass fluxes in evaporative groundwater-source lakes. Symp. Sedimentology and geochemistry of modern and ancient saline lakes. SEPM Special Publication 50, 33-50.

Ewing, B., Green, P., 1998. Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res. 8, 186-194.

36 Ewing, B., Hillier, L., Wendl, M. C., Green, P., 1998. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 8, 175- 185.

Galat, D. L., Lider, E. L., Vigg, S., Robertson, S. R., 1981. Limnology of a large, deep, North American terminal lake, Pyramid Lake, Nevada, U.S.A. Hydrobiologia 81-82, 281-317.

Gordon, D., Abajian, C., Green, P., 1998. Consed: A graphical tool for sequence finishing. Genome Res. 8, 195-202.

LaMontagne, M. G., Michel, F. C., Holden, P. A., Reddy, C. A., 2002. Evaluation of extraction and purification methods for obtaining PCR-amplifiable DNA from compost for microbial community analysis. J. Microbiol. Methods 49, 255- 264.

Lanave, C., Preparata, G., Saccone, C., Serio, G., 1984. A new method for calculating evolutionary substitution rates. J. Mol. Evol. 20, 86-93.

Lane, D. J. 1991. 16s/23s rRNA sequencing. In: Stackerbrandt, E. G., M. (ed.). Wiley, New York.115-175.

Lasek-Nesselquist, E., Bogomolni, A. L., Gast, R. J., Welch, D. M., Ellis, J. C., Sogin, M. L., Moore, M. J., 2008. Molecular characterization of Giardia intestinalis haplotypes in marine animals: variation and zoonotic potential. Dis. Aquat. Organ. 81, 39-51.

Li, J. G., Philp, R. P., Pu, F., Allen, J., 1996. Long-chain alkenones in Qinghai Lake sediments. Geochimica Et Cosmochimica Acta 60, 235-241.

Liu, W. G., Liu, Z. H., Fu, M. Y., An, Z. H., 2008. Distribution of the C37 tetra- unsaturated alkenone in Lake Qinghai, China: A potential lake salinity indicator. Geochimica Et Cosmochimica Acta 72, 988-997.

Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar, Buchner, A., Lai, T., Steppi, S., Jobb, G., Forster, W., Brettske, I., Gerber, S., Ginhart, A. W., Gross, O., Grumann, S., Hermann, S., Jost, R., Konig, A., Liss, T., Lussmann, R., May, M., Nonhoff, B., Reichel, B., Strehlow, R., Stamatakis, A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K. H., 2004. ARB: a software environment for sequence data. Nucleic Acids Res. 32, 1363-1371.

Maddison, D. R., Maddison, W. P. (2000) MacClade version 4: Analysis of phylogeny and character evolution. Sunderland, Massachusetts, Sinauer Associates.

37 Marlowe, I. T., Green, J. C., Neal, A. C., Brassell, S. C., Eglinton, G., Course, P. A.,

1984. Long-Chain (n-C37-C39) alkenones in the prymnesiophyceae. Distribution of alkenones and other lipids and their taxonomic significance. British Phycological Journal 19, 203-216.

McGowan, S., Ryves, D. B., Anderson, N. J., 2003. Holocene records of effective precipitation in West Greenland. Holocene 13, 239-249.

Medlin, L., Elwood, H., Stickel, S., Sogin, M., 1988. The characterization of enzymatically amplified eukaryotic 16S-like rRNA coding regions. Gene 71, 491-499.

Medlin, L. K., Saez, A. G., Young, J. R., 2008. A molecular clock for coccolithophores and implications for selectivity of phytoplankton extinctions across the K/T boundary. Marine Micropaleontology 67, 69-86.

Morgan, D. L., 1947. The Great Salt Lake. University of Utah Press, Salt Lake City.

Müller, P. J., Kirst, G., Ruhland, G., von Storch, I., Rosell-Mele, A., 1998. k' Calibration of the alkenone paleotemperature index U37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60 °S). Geochimica Et Cosmochimica Acta 62, 1757-1772.

Overmann, J., Beatty, J. T., Hall, K. J., Pfennig, N., Northcote, T. G., 1991. Characterization of a Dense, Purple Sulfur Bacterial Layer in a Meromictic Salt Lake. Limnol. Oceanogr. 36, 846-859.

Pearson, E. J., Juggins, S., Farrimond, P., 2008. Distribution and significance of long- chain alkenones as salinity and temperature indicators in Spanish saline lake sediments. Geochimica Et Cosmochimica Acta 72, 4035-4046.

Pienaar, R. N., 1980. Observations on the structure and composition of the cyst of Prymensium (Prymnesiophyceae). Proceedings of the Electron Microscopy Society of Southern Africa 10, 73-74.

Prahl, F. G., Muehlhausen, L. A., Zahnle, D. L., 1988. Further evaluation of long- chain alkenones as indicators of paleoceanographic conditions. Geochimica Et Cosmochimica Acta 52, 2303-2310.

Prahl, F. G., Wakeham, S. G., 1987. Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330, 367-369.

Pruesse, E., Quast, C., Knittel, K., Fuchs, B. M., Ludwig, W. G., Peplies, J., Glockner, F. O., 2007. SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res. 35, 7188-7196.

38

Rodriguez, R., Oliver, J. L., Marin, A., Medina, J. R., 1990. The general stochastic model of nucleotide substitution. J. Theor. Biol. 142, 485-501.

Ronquist, F., Huelsenbeck, J. P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574.

Rontani, J. F., Beker, B., Volkman, J. K., 2004. Long-chain alkenones and related compounds in the benthic haptophyte Chrysotila lamellosa Anand HAP 17. Phytochemistry 65, 117-126.

Rosell-Melé, A., 1998. Interhemispheric appraisal of the value of alkenone indices as temperature and salinity proxies in high-latitude locations. Paleoceanography 13, 694-703.

Schloss, P. D., Handelsman, J., 2005. Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl. Environ. Microbiol. 71, 1501-1506.

Schulz, H.-M., Schöner, A., Emeis, K.-C., 2000. Long-chain alkenone patterns in the Baltic sea--an ocean-freshwater transition. Geochimica Et Cosmochimica Acta 64, 469-477.

Sicre, M. A., Marty, J. C., Saliot, A., 1990. n-Alkanes, fatty acid esters, and fatty acid salts in size fractionated aerosols collected over the Mediterranean Sea. J. Geophys. Res. 95, 3649-3657.

Simon, N., Campbell, L., Örnolfsdottir, E., Groben, R., Guillou, L., Lange, M., Medlin, L. K., 2000. Oligonucleotide probes for the identification of three algal groups by dot blot and fluorescent whole-cell hybridization. The Journal of Eukaryotic Microbiology 47, 76-84.

Sun, Q., Chu, G. Q., Liu, G. X., Li, S., Wang, X. H., 2007. Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellosa (Haptophyceae). Organic Geochemistry 38, 1226-1234.

Tavaré, S., 1986. Some probabilistic and statistical problems in the analysis of DNA sequences. Lectures on Mathematics in the Life Sciences 17, 57-86.

Toney, J. L., Huang, Y., Fritz, S. C., Baker, P. A., Grimm, E., Nyren, P., 2009. Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes in the interior United States. Geochimica Et Cosmochimica Acta. 74, 10.1016/j.gca.2009.11.021

39 Valero-Garces, B. L., Kelts, K. R., 1995. A sedimentary facies model for perennial and meromictic saline lakes - Holocene Medicine Lake Basin, South-Dakota, USA. Journal of Paleolimnology 14, 123-149.

k' Versteegh, G. J. M., Riegman, R., de Leeuw, J. W., Jansen, J. H. F., 2001. U37 values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations. Organic Geochemistry 32, 785-794.

Volkman, J. K., Barrett, S. M., Blackburn, S. I., Sikes, E. L., 1995. Alkenones in Gephyrocapsa oceanica - Implications for Studies of Paleoclimate. Geochimica Et Cosmochimica Acta 59, 513-520.

Volkman, J. K., Eglinton, G., Corner, E. D. S., Forsberg, T. E. V., 1980. Long-Chain Alkenes and Alkenones in the Marine Coccolithophorid Emiliania-Huxleyi. Phytochemistry 19, 2619-2622.

Yurkova, N., Rathgeber, C., Swiderski, J., Stackebrandt, E., Beatty, J. T., Hall, K. J., Yurkov, V., 2002. Diversity, distribution and physiology of the aerobic phototrophic bacteria in the mixolimnion of a meromictic lake. FEMS Microbiol. Ecol. 40, 191-204.

Zink, K. G., Leythaeuser, D., Melkonian, M., Schwark, L., 2001. Temperature dependency of long-chain alkenone distributions in Recent to fossil limnic sediments and in lake waters. Geochimica Et Cosmochimica Acta 65, 253- 265.

40 Table 1. Lake locations, maximum depth and salinity.

Name Location Lat Long Max. depth (m) Salinity (ppt) AndersonReference et al., 2004; BrayaSø Greenland 66.99 -51.01 22.4 0.9-1.9 D'Andrea et al., 2006 Donovan, 1994; Zink et al., 2001; Toney et Brush Lake Montana, USA 48.60 -104.11 18 3 al., 2009 Clear Lake North Dakota, USA 47.86 -100.10 1.7 5 Toney et al., 2009 http://www.homepages. ucl.ac.uk/~ucfasdt/Tibe GaHai Tibetan plateau, China 37.13 97.57 10.5 - t/index.htm Great Salt Lake Utah, USA 41.17 -112.58 10 270 Morgan, 1947 http://www.homepages. ucl.ac.uk/~ucfasdt/Tibe Keluke Hu Tibetan plateau, China 37.28 96.89 8 - t/index.htm Lake George North Dakota, USA 46.74 -99.49 18.8 10 Toney et al., 2009 Li et al., 1996; Liu et Lake Qinghai Tibetan plateau, China 36.53 99.60 27 15 al., 2008 Overmann et al., 1991; Mahoney Lake British Columbia, Canada 49.00 -119.00 17.8 13240 Yurkova et al., 2002 Valero-Garces and Kelts, 1995; Toney et Medicine Lake South Dakota, USA 44.82 -97.35 8.3 9 al., 2009 Moon Lake North Dakota, USA 46.86 -98.16 8.2 3 Toney et al., 2009 Pyramid Lake Nevada, USA 40.00 -119.50 103 5 Galat et al., 1981 Skoal Lake North Dakota, USA 47.92 -101.47 1.8 8 Toney et al., 2009 http://www.homepages. ucl.ac.uk/~ucfasdt/Tibe Tso Ur Tibetan plateau, China 31.48 91.52 8.4 - t/index.htm Upper Murray Elesmere Island, Canada 81.33 -69.50 82 0.04-0.06 Besonen et al., 2008

41 Table 2. Primers used for SSU rRNA gene amplification.

Universal eukaryotic primer 5'!3' Tm (°C) Reference 1391R GAC GGG CGG TGT GTR CA 57 Lane, 1991 EukA 5'F AC CTG GTT GAT CCT GCC AGT 57 Medlin et al., 1988 EukB 3'R C TTC TGC AGG TTC ACC TAC 53 Medlin et al., 1988 Haptophyte-specific primers Prym429F GCG CGT AAA TTG CCC GAA 65 Coolen et al., 2004 Prym887R (aka PRYM02) GGA ATA CGA GTG CCC CTG AC 62 Simon et al., 2000 Greenland haptophyte-specific GL625R G CTA AGA GTA GGC GCG GTC TTT GGG AGG 74 this work

42 Table 3. Percent OTU representation in lake sediment clone libraries. (A)

Surface sediment distribution; (B) Greenland Lake BrayaSø downcore sediment relative OTU distribution. Approximate age given (Anderson and Leng, 2004).

(+) denotes presence of OTU but a low confidence value for percentage.

A. Group I Group II Sequences OTU 1 OTU 2 OTU 3 OTU 4 OTU 5 OTU6 OTU 7 OTU 8 recovered Symbol BrayaSo, Greenland (total) 50% 1% 1% 2% 47% - - - 187 BS Clear lake, ND, USA ------+ <5 CL Great Salt Lake, UT, USA ------+ - <5 GS Keluke Hu downcore, TP, CN - - - - 60% - 20% 20% 5 KE Lake George, ND, USA ------53% 47% 62 LG Medicine Lake, SD, USA ------40% 60% 5 MD Pyramid Lake, NV, USA - - - - 29% - 57% 14% 14 PY Skoal Lake, ND, USA 6% - - - - - 63% 31% 16 SK Tso Ur downcore, TP, CN 37% 63% ------94 TUd Tso Ur surface, TP, CN - - - - - 4% 38% 58% 24 TUs Upper Murray Lake, EI, CA 75% - - - 25% - - - 24 UM

B. Sequences OTU 1 OTU 2 OTU 3 OTU 4 OTU 5 recovered Age (14C yrs BP) BrayaSo core 24 cm 53% - - 1% 46% 81 1184 BrayaSo core 40 cm 51% 1% - 1% 47% 77 2039 BrayaSo core 50 cm 56% - - - 44% 73 2574 BrayaSo core 70 cm 26% - 3% 1% 71% 147 3643 BrayaSo core 88 cm - - - - + <5 4606 BrayaSo core 110 cm 45% - - 6% 49% 83 5783

43 Table 4: Uncorrected distance matrices of pairwise comparisons among sequences. (A) Representative OTU sequences and Greenland BrayaSø water sequence; (B) Reference taxa and Greenland BrayaSø water sequence. Numbers are % difference.

OTU 1 BrayaSø water OTU 2 OTU 3 OTU 4 OTU 5 OTU 6 OTU 7 OTU 8 OTU 1 - 0.2 4.8 2.8 2.6 0.4 6.7 4.3 5.2 BrayaSø water - - 4.6 2.6 2.4 0.2 6.5 4.1 5 OTU 2 - - - 2.4 6.5 4.3 10.2 8 9.1 OTU 3 - - - - 4.5 2.4 8.7 6.3 7.1 OTU 4 - - - - - 2.2 8.4 6 6.9 OTU 5 ------6.3 3.9 4.8 OTU 6 ------3.2 5 OTU 7 ------2.2 OTU 8 ------

44 Table 5. Alkenone trends among haptophyte algae. (A) Reference taxa; (B) lakes reported in this study.

Phylogenetic K' * A. Coastal/Lacustrine C37/C38 C38 MeK C40 group U37 calibration % C37:4 Chysotila lamellosa (marine) > 2 absent present Group II Uk37' =0.01T+0.0033a Uk37' = 0.0011T2 - Chysotila lamellosa (lacustrine) > 2 absent present Group II 0.0157T + 0.1057b Isochrysis galbana > 3 absent present Group II Uk37' = 0.00932 - 0.0413c Ace Lake, Antarctica > 2 absent present Group II NRd Lake BrayaSo, Greenland 1.3 present absent Group I Uk37' = (T + 5.4)/155.2e 65

Marine Emiliania huxleyi 1-2 present absent Group III Uk37' = 0.033T + 0.043f Gephyrocapsa oceanica < 1 present absent Group III Uk37' = 0.049T - 0.52g

B. This work Clear lake, North Dakota > 2 - - Group II - Great Salt Lake, Utah 1.9 present trace? Group II 31 Keluke Hu downcore, Tibetan Plateau 0.9 - trace Group I, II - Lake George, North Dakota 2.0 present trace Group II 55 Medicine Lake, South Dakota 3.8 absent trace? Group II 67 Pyramid Lake, Nevada 1.9 present trace? Group I, II 22 Skoal Lake, North Dakota 2.2 - - Group I, II - Tso Ur surface, Tibetan Plateau 1.7 present trace Group II 30 Tso Ur downcore, Tibetan Plateau 1.8 present trace Group I 27 Upper Murray Lake, Ellesmere Island 2.4 present absent Group I 63

* %C37:4 = C37:4/(C37:2 + C37:3 + C37:4) a. Conte et al., 1994; Rontani et al., 2004 b. Sun et al., 2007 c. Marlowe et al., 1984; Conte et al., 1994; Versteegh et al., 2001 d. Coolen et al., 2004 e. D'Andrea and Huang, 2005; D'Andrea, 2008 f. Volkman et al., 1980; Chu et al., 2005; Liu et al., 2008; Prahl and Wakeman, 1987 g. Volkman, 1995

45

Figure 1. Map of alkenone-containing lakes. Squares denote previously reported lakes, circles denote lakes analyzed in this study.

46 Figure 2. A consensus Bayesian phylogenetic tree depicting 18S rRNA gene- inferred relationships among haptophyte algae. An asterisk (*) indicates posterior probability values of 1.00; all other values as shown. Bold designates sequences from this study. Operational Taxonomic Units (OTUs) were defined at a 95 % similarity level. Order classification after de Vargas et al. (2007), with number of sequences per order as indicated. The evolutionary distance for the number of changes per site is represented by the scale bar. GenBank accession numbers follow species and sequence names. Lakes with given OTU representation shown in brackets; abbreviations as shown in Table 3.

47 Figure 3. A consensus Bayesian phylogenetic tree depicting 18S rRNA gene- inferred relationships among haptophyte algae using only full-length sequences.

Note improved support for the placement of the Greenland BrayaSø sequence.

Operational Taxonomic Units (OTUs) defined at a 95 % similarity level. An asterisk (*) indicates posterior probability values of 1.00; all other values as shown. The evolutionary distance for the number of changes per site is represented by the scale bar. GenBank accession numbers follow species names and sequence names.

48

Figure 4. Lake sediment alkenone signatures. A) Upper Murray Lake Group I example, B) Medicine Lake Group II example, C) Pyramid lake with both

Group I and Group II sequences.

49

CHAPTER 3

COMPARATIVE MOLECULAR MICROBIAL ECOLOGY OF A SPRING

HAPTOPHYTE BLOOM IN A GREENLAND ARCTIC OLIGOSALINE LAKE

SUSANNA THEROUX1,2

LINDA AMARAL-ZETTLER1,2,3

YONGSONG HUANG1

1 BROWN UNIVERSITY, Department of Geological Sciences

2 MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

3 BROWN UNIVERSITY, Department of Ecology and Evolutionary Biology

In review:

Frontiers in Microbiology

50

Abstract

The Arctic is highly sensitive to increasing global temperatures and is projected to experience dramatic ecological shifts in the next few decades.

Oligosaline lakes are common in arctic regions where evaporation surpasses precipitation, however these extreme microbial communities are poorly characterized.

Many oligosaline lakes, in contrast to freshwater ones, experience annual blooms of haptophyte algae that generate valuable alkenone biomarker records that can be used for paleoclimate reconstruction. These haptophyte algae are globally important, and globally distributed, aquatic phototrophs yet their presence in microbial molecular surveys is scarce. To target haptophytes in a molecular survey, we compared microbial community structure during two haptophyte bloom events in an arctic oligosaline lake, Lake BrayaSø in southwestern Greenland, using high-throughput pyrotag sequencing. Our comparison of two annual bloom events yielded surprisingly low taxon overlap, only 13% for bacterial and 26% for eukaryotic communities, which indicates significant annual variation in the underlying microbial populations.

Both the bacterial and eukaryotic communities strongly resembled high-altitude and high-latitude freshwater environments. In spite of high alkenone concentrations in the water column, and corresponding high haptophyte rRNA gene copy numbers, haptophyte pyrotag sequences were not the most abundant eukaryotic tag, suggesting that sequencing biases obscured relative abundance data. With over 170 haptophyte tag sequences, we observed only one haptophyte algal Operational Taxonomic Unit, a prerequisite for accurate paleoclimate reconstruction from the lake sediments. Our

51 study is the first to examine microbial diversity in a Greenland lake using next generation sequencing and the first to target an extreme haptophyte bloom event. Our results provide a context for future explorations of aquatic ecology in the warming arctic.

52 1. Introduction

Oligosaline lakes (salinity 0.5-5 ppt) develop in polar regions near ice sheets where evaporation exceeds precipitation and provide a unique habitat apart from the more common glacially-derived freshwater lakes. These high latitude lakes serve as sensitive indicators of the ecosystem response to global climate change (Marchetto,

2004; Quayle et al., 2002) as their low salinity reflects small changes in hydrological balance. In the past decade alone, southwestern Greenland has undergone marked warming, and major warming is predicted for the future (Bennike et al., 2010).

Microbial surveys targeting the 18S ribosomal RNA (rRNA) gene have revealed previously unknown diversity in microbial eukaryotes lineages such as cryptomonads, katablepharids, dinoflagellates and Perkinsea (Logares et al., 2007; Slapeta, Moreira, and Lopez-Garcia, 2006; Shalchian-Tabrizi et al., 2011). However, haptophyte algae have been largely absent from these studies, potentially the result of naturally low haptophyte abundances in the environments selected such as deep sea habitats or anoxic lakes (Stoeck et al., 2009; Stoeck et al., 2010; Edgcomb et al., 2011;

Pawlowski et al., 2011; Shalchian-Tabrizi et al., 2011). The GC-rich haptophyte genomes may also hinder amplification reactions that use universal primer sets

(Moon-van der Staay et al., 2001; Liu et al., 2009; Stoeck et al., 2010). In this study, we targeted the haptophyte-rich waters of an arctic oligosaline lake spring bloom event to shed light on the microbial diversity of these unique ecosystems.

Lake BrayaSø in southwestern Greenland experiences a seasonal haptophyte bloom approximately two weeks after ice off (D’Andrea et al., 2011). These haptophyte blooms result in exceptional abundances of alkenones in BrayaSø

53 sediments (82 mg/g total organic carbon, D’Andrea et al., 2005) that provide the first quantitative temperature record for the past five thousand years for southwestern

Greenland (D’Andrea et al., 2011). Only a few species of haptophyte algae, in the order Isochrysidales, produce alkenone lipids. These species and their alkenone lipids have been extensively studied in marine environments, where alkenones are preserved in marine sediments as a record of sea surface temperature back through time (Brassell et al, 1986; Conte et al, 2006; Marlowe et al., 1984; Prahl and

Wakeham, 1987; Volkman et al, 1980; Muller et al, 1998). The endeavor to extend this alkenone-based proxy to the continents has resulted in pan-continental surveys of lake sediments and waters for alkenone-producing haptophyte algae. The use of haptophyte-specific primers targeting environmental DNA has revealed considerable diversity in lake-dwelling haptophytes (Coolen et al., 2004; D'Andrea et al., 2006;

Theroux et al., 2010). However, these haptophytes are largely absent in molecular surveys targeting universal genes and microbial diversity studies of haptophyte blooms are nonexistent.

Nutrient loading and seasonal irradiance levels are known to trigger marine and estuarine haptophyte blooms (Tyrrell and Merico, 2004) that occur across latitudes in both cold and coastal regions (Brown and Yoder, 1994). Increasing global temperatures will result in longer ice-free periods in arctic lakes and an increase in runoff from thawing tundra catchments, undoubtedly affecting the annual haptophyte bloom events. Given the significance of haptophyte algae in aquatic ecosystems, their absence in previous datasets, and the desire to anticipate their future response to global climate change, the objectives of our study were two fold: (1) to sequence a

54 haptophyte-rich environment with a universal molecular approach; and (2) to produce a benchmark species survey for an arctic oligosaline lake during the spring bloom.

Using high-throughput pyrotag sequencing, we targeted both bacterial and eukaryotic communities from two separate years to evaluate the consistency of the bloom- associated microbial populations. Our study provides an important baseline to contrast future BrayaSø microbial community change during its ice-free period in anticipation of a warmer Arctic possessing longer ice-free periods.

2. Materials and Methods

2.1 Site description

The Kangerlussuaq region of Southwestern Greenland lies at the head of the

Søndre Strømfjord, 150 km from the ocean outlet, and has a series of saline lakes that provide ideal locations of paleoclimate reconstruction due to their anoxic hypolimnions and excellent preservation of lake sediments (Figure 1). Lake BrayaSø

(66.99°N, -51.01°W) is a meromictic, oligosaline lake (salinity 2-3) that has an ice cap from September to late May. The surface area of Lake BrayaSø is approximately

72 hectares, with a maximum depth of 24 m. The dominant salts are NaCO3, NaHCO3 and MgHCO3, which are delivered to the lakes via aeolian transport from nearby sand sheets and input from erosion within the lake catchments (Anderson et al., 2001). The bedrock in the region is granodioritic gneiss with occasional ultrabasic intrusions

(Heggen et al., 2001). The climate is low-Arctic continental with > 500 mm/yr of precipitation and a mean annual temperature at Kangerlussuaq of -6°C (Heggen et al.,

2001). The lake is calcium-depleted relative to fresh lakes of the region due to CaCO3

55 precipitation, and the dominant cations are Na+> Mg2+ > K+> Ca2+ (D’Andrea 2008).

Total nitrogen is approximately 803 μg/L and total phosphorus is approximately 9

μg/L (Brutemark et al., 2006). Dissolved organic carbon is approximately 90 mg/L

(Anderson et al. 2009). Alkenone lipids are present in the sediments of Lake BrayaSø

(D’Andrea et al., 2005) and sediment trap data indicated the annual haptophyte bloom in the lake occurs approximately two weeks after ice-off (D’Andrea, 2008).

2.2 Water column

We selected samples for pyrosequencing based upon chlorophyll and alkenone concentrations in the water column, choosing the 10-m depth for 2007 and 4-m depth from 2009 (Figure 2). For 2007, 10-m corresponded to the peak in chlorophyll and alkenone concentrations. For 2009, 4-m depth was slightly above the chlorophyll maximum and coincided with the alkenone concentration peak (Figure 1). The June average temperature for each sampling day in 2007 and 2009 was 11.2°C and 9.9°C, respectively (Figure 3A). The average daily air temperature for June 2007 was

10.7°C and for June 2009 was 10°C (Figure 3B).

2.3 Water sampling

The bloom event occurred in BrayaSø, Greenland in June of 2007 and 2009. For both years, we analyzed a sample collected during the first week of the haptophyte bloom.

We collected geochemical data using a YSI Sonde (Ohio, USA) equipped with probes to measure temperature, conductivity, dissolved oxygen, and chlorophyll a fluorescence. At 1-m intervals, we collected water with a Van Dorn water sampler

56 and preserved these samples for alkenone and genomic DNA analysis. For alkenone analysis, we filtered one liter of water onto a pre-combusted (550°C) GF/F 0.7μm, 47 mm glass filter, and kept it frozen until analysis. For DNA analysis, we filtered a separate liter of lake water onto a 0.2μm SterivexTM filter (Millipore, Billerica, MA,

USA), flooded the filter with Puregene lysis buffer (Qiagen, Valencia, CA, USA), and froze it at -20°C until processing. We selected samples for sequencing based on maximum alkenone concentrations.

2.4 Lipid analysis

Alkenone extraction was after D'Andrea and Huang (2005). Alkenone samples and

DNA samples were sourced from the same water sample. We freeze-dried and homogenized samples manually. We extracted samples with 9:1 Dichloromethane

(DCM):Methanol (MeOH) using an Accelerated Solvent Extractor ASE200 (Dionex,

Sunnyvale, CA, USA). Extracts were separated into acid and neutral fractions using a solution of DCM, Isopropyl alcohol 2:1 (v/v). The neutral fraction was further separated into aliphatic (hexane elution), ketone (DCM), and alcohol (ethyl acetate:hexane 1:3) fractions using a flash silica gel column. The DCM fraction was analyzed using an Agilent 6890plus Gas Chromatograph Flame Ionization Detector

(GC-FID) (Santa Clara, CA) for quantification. Chromatograms were compared to previously reported alkenone standards and their GC retention times (de Leeuw et al.,

1980; Marlowe et al., 1984). Alkenone concentrations were determined from GC-FID

analysis of the ketone fractions based on an internal C36 alkane standard.

57 2.5 DNA extraction

We extracted SterivexTM filters using a Qiagen Puregene Cell Kit (Venlo,

Netherlands) according to the manufacturer’s instructions. Genomic DNA was polyethylene glycol (PEG) purified (LaMontagne et al., 2002) to remove proteins and other contaminants that inhibit PCR reactions. DNA was suspended in PEG at 4°C overnight, centrifuged, and the pellet rinsed with ethanol. The DNA was resuspended in DNA hydration solution (Qiagen). We quantified total extracted genomic DNA yields using a NanoDrop nucleic acid spectrophotometer (Thermo Scientific,

Wilmington, DE) to ensure they were RNA-free.

2.6 Quantitative polymerase chain reaction

Purified DNA extracts were also subjected to real-time quantitative polymerase chain reaction (qPCR) to gauge haptophyte cell concentrations with depth and ensure that the sample selected for sequencing was at the point of highest haptophyte cell concentration in the water column. We performed the qPCR reaction using 18S rRNA gene haptophyte-specific primers Prym-429F (5’-GCG CGT AAA TTG CCC GAA-

3’; Tm = 65°C), and Prym-887R (5’-GGA ATA CGA GTG CCC CTG AC-3’; Tm =

62°C) (Coolen et al., 2004; Simon et al., 2000). These primers yield an amplicon that is approximately 463bp in size. These primers have previously been screened for specificity: the forward primer Prym-429F is 100% specific for Haptophyta order

Prymnesiales and matched 93% of orders Coccosphaerales, Isochrysidales,

Prymnesiales, the genus Pleurochrysis, as well as unclassified haptophytes (Coolen et al., 2004). The reverse primer is specific to Prymnesiophyceae (Simon et al., 2000).

58 We further confirmed primer specificity using the ARB probematch tool in SILVA

ARB database v111 (Pruesse et al., 2007). This SSU reference database contains

739,633 high quality 16S/18S rRNA gene sequences. The Prym429F primer matched

72% of full-length 18S rRNA gene haptophyte sequences with two mismatches, and no non-haptophyte sequences. The Prym88R primer returned 99.3% of haptophyte sequences with one mismatch and no non-haptophyte sequences.

The qPCR reactions were run in triplicate, including a no-template control, on an Applied Biosystems StepOnePlusTM Real-Time PCR System (Foster City,

California), using a SYBR Green I assay. We also ran a positive control of Isochrysis galbana DNA extracted from a culture with cell concentrations at 1.5x106 cells/ml.

The Cq for each sample had a deviation of less than 0.5. Each 20 µl reaction contained 7.2 µl of sterile water, 10 µl of KAPA SYBR® FAST Universal 2X qPCR

Master Mix (Woburn, MA), 0.4 µl each of the forward and reverse primers (0.2µM) and 2 µl of template DNA. Template DNA ranged in concentration from 2-10ng/µl

The qPCR cycling program was after Coolen et al. (2009) and consisted of 38 cycles of denaturation at 94°C for 30s, annealing at 62°C for 40s, primer extension at 72°C for 60s, a photo step of 80°C for 20s. We used between 101 and 106 copies (ten-fold dilution series) of linearized plasmids containing the 18S rRNA gene of Isochrysis galbana CCMP1323 as the external standard to calibrate the copy numbers of haptophyte RNA genes in the BrayaSø water samples. Our standard curve was established using four points of the diluted standard, with an R2 value of >.999 and slope of –3.991. Our reaction efficiency was 78.1%. We used StepOne Software version 2.2 (Applied Biosystems) to analyze our results. These conditions are

59 reported in accordance with the Minimum Information for Publication of Quantitative

Real-Time PCR Experiments (MIQE) guidelines (Bustin et al., 2009).

2.7 Pyrosequencing

We performed genomic DNA amplifications using eukaryotic and bacterial-specific primers targeting the V9 (Amaral-Zettler et al., 2009) or V6-V4 (Morrison and Sogin, in prep.) regions, respectively. Eukaryotic sequences were generated using a Genome

Sequencer FLX (Roche, Switzerland) with the GS-LR70 long-read sequencer kit at the Marine Biological Laboratory Keck Sequencing Facility. Amplifications and sequencing for eukaryotic sequences were after Amaral Zettler et al. (2009). We sequenced the V6-V4 hypervariable region of the bacterial 16S rRNA gene using bacterial primers 515F and 1046R on a Roche GS FLX pyrosequencer using GS FLX

Titanium Series reagents (Roche Diagnostics, Basel, Switzerland) following manufacturer’s protocols. Sequences were trimmed and screened for quality after

Huse et al. (2007). To assign taxonomy to the remaining quality-controlled tags, we used the Global Alignment for Sequence Taxonomy (GAST) algorithm (Huse et al.,

2008). Tag sequences were grouped into Operational Taxonomic Units using SLP-

PWAL (refer to Huse et al. 2010), with bacterial sequences clustered at 3% and eukaryotic sequences clustered at 6%. Venn diagrams were constructed using

BioVenn (Hulsen et al., 2008). Bacterial diversity estimates were calculated using

EstimateS v8.2.0 (Colwell 2005) and CatchAll (Bunge, 2011). All sequences have been deposited in the NCBI Sequence Read Archive (SRA) under the SRA number

SRA059384, and are MIMARKS compliant (Yilmaz et al., 2011).

60 3. Results and Discussion

3.1 Water column

Both 2007 and 2009 samples were selected from the first week of the two- week haptophyte bloom. In both years, secchi depth was 5 m indicating the photic zone terminated at approximately 10-12.5 m depth. The alkenone peak in 2007 corresponded to the oxycline at 10m depth, whereas the alkenone peak in 2009 was at the thermocline.

Quantitative PCR analysis confirmed that the water sample from peak alkenone depth corresponded to peak haptophyte cell numbers (Table 1, Figure 2). The structure of the water column between the two years was markedly different, with chlorophyll peaking at 10 m depth in 2007 and about 6.5m depth in 2009. Alkenone concentrations peaked at 15 μg/L in 2007 and about 59 μg/L in 2009.

Correspondingly, rRNA gene copies peaked at 2146 copies/ml in 2007 and 9898 copies/ml in 2009 (Figure 1). This equates to approximate cellular alkenone concentrations of 6-7 ng per cell if the 18S rRNA gene copies occur singularly or 3-

3.5 ng/cell if there are two copies of the 18S rRNA gene in these haptophytes. This is on par with previously observed cellular alkenone concentrations in lacustrine haptophytes of 0.009 to 2 pg/cell (Versteegh et al., 2001; Marlowe et al., 1984).

Given that Lake BrayaSø has the highest sedimentary concentrations of alkenones ever reported (D’Andrea and Huang, 2005), the high cellular concentrations as estimated by our qPCR analysis is not surprising. Our results also agree with observations by Boere et al. (2011) that alkenone concentrations can serve as a proxy for haptophyte cell numbers.

61

3.2 Bacterial community diversity

A total of 6,409 bacterial OTUs were observed between 2007 and 2009

(Figure 4A). The two years had comparable OTU yields, 2883 from 2007 and 2727 from 2009, yet shared only 13% of their OTUs. This overlap in OTUs was surprisingly low, considering reports from other arctic lake surveys demonstrating up to 73% overlap in bacterial community membership (Crump et al., 2003). When singletons were ignored, this overlap increased to 44% (Figure 4B), indicating a third of the taxon differences came from the rarest members of the community. Our study revealed much greater bacterial diversity at the phylum level, 25 phyla, than previous studies from freshwater and oligosaline lakes on the Tibetan plateau that identified only 13 phyla (Liu et al., 2010). The estimates of alpha diversity of the bacterial community generated species richness estimates with overlapping confidence bounds

(Table 2), demonstrating that our similar OTU yields reflected the similar alpha diversity or richness during the two years.

The Morisita-Horn index of similarity, an abundance-based distance measure of beta diversity, was 0.848 (complete overlap = 1) (Table 2) indicating the most abundant species were present in both 2007 and 2009. Of the most abundant bacterial

OTUs (Table 3A), the notable difference between communities in 2007 and 2009 was the presence of the sulfur-oxidizing bacteria in 2009 (Thiomicrospira,

Sulfurovum, Sulfuricurvum) and fewer Flavobacteria in 2009. This flavobacterial

OTU matched environmental sequences from freshwater environments, including

100% sequence identity to a bacterium isolated during a spring phytoplankton bloom

62 in Lake Zurich (Eckert et al., 2011) and bacteria from lakes on the Tibetan plateau

(Zhang, R. and Liu, W.-T. unpublished, GenBank HM128691).

While Lake BrayaSø is oligosaline, its bacterial community resembled previously reported freshwater environments in addition to high altitude environments. The bacterial OTUs were dominated by Actinobacteria, which are known to occur ubiquitously in terrestrial and aquatic ecosystems (Embley and

Stackebrandt, 1994) and can dominate lake epilimnia (Newton et al., 2011). The most abundant actinobacterial OTU matched environmental sequences from Lake Taihu

(China) and other freshwater lakes with 100% identity. The second most abundant phylum represented, the beta-proteobacteria, occurs more commonly in freshwater environments than marine (Nold et al., 1998) and represents the most abundant bacteria in glacial meltwater communities (Cheng and Foght, 2007). Overall, the most abundant bacterial taxa (Table 3A) matched sequences from other freshwater environments, and resembled that of high-altitude lakes from the Tibetan plateau in the abundance of Actinobacteria, alpha- and beta- Proteobacteria (Xing et al., 2009;

Liu et al., 2010). Lakes at high altitude experience similar environmental pressures as lakes at high latitude, including oligotrophy, low temperature, and high UV radiation in the surface waters; the similarity in their bacterial communities suggests these particular phyla can withstand harsh environmental conditions across latitudes.

Toolik Lake in Alaska experiences an increase in primary and bacterioplankton production in the first month of spring as melting snow increases organic matter and nutrient transport into the lake and allows for an increase in sunlight reaching the water column (Hobbie et al., 1983; Crump et al., 2003). A

63 similar trend is observed on the western shelf of the Antarctic peninsula, where seasonal melting dictates irradiance levels, mixed layer depth, and organic carbon availability (Montes-Hugo et al., 2010), with an increase in primary production resulting in an increase in bacterial production. Given the increasing global temperatures, we anticipate an increased supply of organic matter into Lake BrayaSø and thus an increase in bacterioplankton production. An increasing freshwater input as a result of melting arctic tundra may affect the local hydrological balance enough to freshen Lake BrayaSø and shift the microbial community further towards one of a more freshwater composition.

3.3 Eukaryotic community diversity

Previous work in Lake BrayaSø identified only 11 eukaryotic phyla

(Brutemark et al., 2006); using high throughput sequencing we were able to identify 9 times more phyla, including picoplankton that were undetectable with the previous visual identification methods (Table 3B). A total of 97 eukaryotic OTUs were observed between 2007 and 2009, with an overlap between the two years of only 26%

(Figure 4C, Table 2). When singletons were ignored, this number increased to 41%

(Figure 4D). The eukaryotic community was dominated by diatom and ciliate OTUs in spite of the presence of a haptophyte “bloom”. Diatoms are known to have high copy numbers of their 18S rRNA genes, which may be the cause of the high abundance of their OTUs (Zhu et al., 2005; Not et al. 2009), as are alveolates which range up to 9,000 copies per cell (Prescott, 1994). Haptophyte 18S rRNA gene copy numbers are estimated at 2-3 copies per cell (Zhu et al., 2005) and our qPCR analysis

64 using an Isochrysis galbana standard yielded approximate 18S rRNA gene copy number at 1 copy per cell (Table 1). Given the high ciliate and diatom tag sequences, these patterns in eukaryotic community structure likely reiterate a cautionary note on the interpretation of abundance data for 18S rRNA gene studies, although these concerns can be minimized when comparing intra-species abundances.

The eukaryotes present in BrayaSø were typical of freshwater meso- and eutrophic environments. Diatom-related OTUs were the most abundant tags we recovered in both 2007 and 2009 (Table 3B). The most abundant diatom OTU represented 31% of all eukaryotic tag sequences, and shared 100% sequence identity with araphid diatoms from fresh and brackish water. The second most abundant OTU in 2007 was assigned a ciliate taxonomy that matched environmental sequences from floodplain soil and an ephemeral pond to 96% and 95%, respectively. In contrast, the second most abundant tag in 2009 matched a metazoan, and shared 100% sequence identity with the copepod Leptodiaptomus moorei (GenBank AY339154). This metazoan was notably absent in the 2007 dataset. Other OTUs present in > 10 fold higher abundance in 2009 versus 2007 included Chlamydomonas, an unidentified environmental ciliate, an alveolate, and a chrysophyte (Table 3B). Of the most abundant eukaryotes, an unidentified ciliate had the greatest average GAST distance of 0.1668 (Table 3B) and is likely a novel species. The similarity of protistan lineages to other freshwater environments confirms the results from other studies increasingly showing the distinction between marine and freshwater communities

(Logares et al., 2009).

65 Haptophytes comprised only 3% of the eukaryotic community in 2007, and

0.4% in 2009 (Table 3B). Despite high levels of recorded alkenones at the depth of sampling, we recovered only one type of haptophyte OTU in both years albeit in much greater abundance in 2007 than 2009 (Table 3B). The most abundant tag in this

OTU cluster shared 100% identity with a previously sequenced 18S rRNA gene from the BrayaSø water column (GenBank HQ446272; (Theroux et al., 2010)). We detected identical haptophyte V9 tag reads to this BrayaSø OTU from Toolik Lake,

Alaska (Crump et al., 2012) and Plum Island, Massachusetts (Amaral-Zettler, personal observation) but nowhere else in the VAMPS global database

(http://vamps.mbl.edu). Singularity of the haptophyte population impacts the ability to use alkenone-derived paleoclimate records from a site; the presence of multiple haptophyte species during a bloom could jeopardize the reliability of the alkenone record. The occurrence of a single haptophyte OTU in both 2007 and 2009 samples is encouraging for the use of this environment as a paleoclimate archive; alkenone- based temperature reconstructions would therefore only require a single temperature calibration for the single haptophyte species present.

The alkenone concentrations at peak alkenone depths in the water column were 15 μg/L in 2007 and 59 μg/L in 2009. Our qPCR results (Table 1) indicated that there were approximately 1700 and 2900 haptophyte cells/ml in the 2007 and 2009 water samples, respectively. Our qPCR results also indicated that Isochrysis galbana has 1-2 copies of its 18 rRNA gene per cell. The occurrence of only 155 and 16 haptophyte tags in 2007 and 2009, respectively, suggested that the haptophyte DNA may have been dwarfed by greater copy number diatom and ciliate 18S rRNA

66 sequences. Empirical obstacles also may have resulted in the low haptophyte tag yield, including primer mismatches or difficulties amplifying GC-rich haptophyte

DNA.

4. Conclusions

Arctic lakes will undoubtedly experience shifts in microbial populations with increasing annual temperatures, prolonged ice-free periods and thawing tundra catchments. This study is the first to examine the microbial community of an artic oligosaline lake using high-throughput sequencing, providing a deeper resolution of the microbial community structure in these rapidly changing arctic environments.

Using high throughput sequencing, we were able to detect greater phylum-richness and new phyla previously unobserved in BrayaSø, the benefit of a molecular versus microscopy-based approach. Even though BrayaSø is an oligosaline lake, both the bacterial and eukaryotic communities resembled other high latitude and high altitude freshwater environments. The low overlap in microbial communities between the

2007 and 2009 samplings suggested large interannual variations in microbial species.

However, the 2009 sample had fewer haptophyte tags but a greater abundance of other phototrophs, suggesting the functional overlap of the eukaryotic communities may be greater than the species overlap. Future studies examining microbial populations throughout the course of a spring bloom event will help resolve these temporal shifts in species abundances and functional roles.

This study is also the first to analyze a haptophyte bloom event using next generation sequencing. We generated fewer haptophyte pyrotag sequences than

67 expected given their alkenone biomarker concentrations in the water column. Our qPCR data confirmed that haptophyte cell numbers peaked with alkenone concentrations, and also showed that high-throughput tag sequences for haptophytes did not correspond well with qPCR counts. This result serves as a reminder that the interpretation of relative abundance data using a tag sequencing approach with eukaryotes must be done so cautiously and that complimentary, haptophyte-specific qPCR provides greater detail of cell abundances. Given the depth of DNA sequencing, and the generation of over 200 haptophyte tags, we are encouraged by the presence of a single haptophyte OTU in Lake BrayaSø, and maintain that this is a worthy location for temperature reconstruction using alkenone-based proxies. Future studies throughout the haptophyte bloom event in BrayaSø will resolve the temporal shifts in microbial communities and will help decipher the communities most susceptible to increasing arctic temperatures.

Acknowledgements

This project was supported by grants from the National Science Foundation (NSF) to

Y. Huang (ARC-0402383), a Brown University SEED grant to Y. Huang and L.

Amaral-Zettler, and an American Association of University Women American

Fellowship to S. Theroux. We thank Melissa Paddock for her assistance with qPCR analyses.

68 References

Amaral-Zettler LA, McCliment EA, Ducklow HW, Huse SM. (2009). A method for studying protistan diversity using massively parallel sequencing of V9 hypervariable regions of small-subunit ribosomal RNA genes. PLoS One 4, e6372.

Anderson NJ, Brodersen KP. (2001). Determining the date of ice-melt for low Arctic lakes along Søndre Strømfjord, southern West Greenland. Geology of Greenland Survey Bulletin 189, 54-58.

Anderson NJ, D'Andrea W, Fritz SC. (2009). Holocene carbon burial by lakes in SW Greenland. Global Change Biology 15:2590-2598.

Bennike O, Anderson NJ, McGowan S. (2010). Holocene palaeoecology of southwest Greenland inferred from macrofossils in sediments of an oligosaline lake. ISME J 43, 787-798.

Boere, A. C., W. I. C. Rijpstra, G. J. de Lange, E. Malinverno, J. S. Sinninghe Damsté, and M. J. L. Coolen. (2011). Exploring preserved fossil dinoflagellate and haptophyte DNA signatures to infer ecological and environmental changes during deposition of sapropel S1 in the eastern Mediterranean, Paleoceanography, 26, PA2204, doi, 10.1029/2010PA001948.

Brassell SC, Eglinton G, Marlowe IT, Pflaumann U, Sarnthein M. (1986). Molecular stratigraphy - a new tool for climatic assessment. Nature 320, 129-133.

Brown, C. and Yoder, J. (1994). Coccolithophorid blooms in the ocean. J. Geophys. Res. 99, 7467–7482.

Brutemark A, Rengefors K, Anderson NJ. (2006). An experimental investigation of phytoplankton nutrient limitation in two contrasting low arctic lakes. Polar Biol 29, 487-494.

Bunge J. (2011). Estimating the number of species with CatchAll. Pac Symp Biocomput 2011, 121-130.

Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley GL, Vandesompele J, and Wittwer CT. (2009) The MIQE Guidelines: Minimum Information for Publication of Quantitative Real-Time PCR Experiments. Clin. Chem. 55, 611-622.

Cheng SM, Foght JM. (2007). Cultivation‐independent and‐dependent characterization of Bacteria resident beneath John Evans Glacier. FEMS Microbial Ecol. 59, 318-330.

69

Colwell, R. K. (2005). EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.2.0. User's Guide and application published at: http://purl.oclc.org/estimates.

Conte MH, Sicre MA, Ruhlemann C, Weber JC, Schulte S, Schulz-Bull D, Blanz T. k' (2006). Global temperature calibration of the alkenone unsaturation index U37 in surface waters and comparison with surface sediments. Geochem Geophy Geosy 7, doi, 10.1029/2005GC001054.

Coolen MJL, Muyzer G, Rijpstra WIC, Schouten S, Volkman JK, Sinninghe Damsté J. (2004). Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth Planet Sci Lett 223, 225-239.

Coolen, MJL. Saenz, JP, Giosan, L, Trowbridge, NY, Dimitrov, P, Dimitrov, D, Eglinton,TI. (2009). DNA and lipid molecular stratigraphic records of haptophyte succession in the Black Sea during the Holocene, Earth Planet Sci Lett 284, 610-621.

Crump BC, Amaral-Zettler LA, Kling GW. (2012). Microbial diversity in arctic freshwaters is structured by inoculation of microbes from soils. ISME J doi, 10.1038/ismej.2012.9.

Crump BC, Kling GW, Bahr M, Hobbie JE. (2003). Bacterioplankton community shifts in an arctic lake correlate with seasonal changes in organic matter source. Appl Environ Microbiol 69, 2253-2268.

D'Andrea WJ, Huang Y. (2005). Long chain alkenones in Greenland lake sediments, Low δ13C values and exceptional abundance. Org Geochem 36, 1234-1241.

D'Andrea WJ, Lage M, Martiny JBH, Laatsch AD, Amaral-Zettler LA, Sogin ML, Huang YS. (2006). Alkenone producers inferred from well-preserved 18S rDNA in Greenland lake sediments. J Geo Res-Biogeo 111, doi, 10.1029/2005JG000121.

D'Andrea WJ. (2008). Development and application of lacustrine alkenone paleothermometry in southwestern Greenland. PhD dissertation, 197. de Leeuw JW, van der Meer FW, Rijpstra WIC, Schenck PA. (1980). On the occurrence and structural identification of long chain unsaturated ketones and hydrocarbons in sediments. Phys Chem Earth 12, 211-217.

Eckert EM, Posch MMST, Eugster B, Pernthaler J. (2011). Rapid successions affect microbial N‐acetyl‐glucosamine uptake patterns during a lacustrine spring

70 phytoplankton bloom. Environ Microbiol 14, 794-806.

Edgcomb V, Orsi W, Bunge J, et al. (2011) Protistan microbial observatory in the Cariaco Basin, Caribbean. I. Pyrosequencing vs Sanger insights into species richness. ISME Journal 5, 1344-1356.

Embley TM, Stackebrandt E. (1994). The molecular phylogeny and systematics of the actinomycetes. Annu Rev Microbiol 48, 257-289.

Haukka K, Kolmonen E, Hyder R, Hietala J, Vakkilainen K, Kairesalo T, Haario H, Sivonen K. (2006). Effect of nutrient loading on bacterioplankton community composition in lake mesocosms. Microb Ecol 51, 137-146.

Hulsen T, De Vlieg J, Alkema W. (2008). BioVenn–a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC genomics 9, 488.

Huse SM, Mark Welch D, Morrison HG, Sogin ML. (2010). Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environ Microbiol 1889-1898.

Huse SM, Huber JA, Morrison HG, Sogin ML, Welch DM. (2007). Accuracy and quality of massively parallel DNA pyrosequencing. Genome Biol 8, R143.

Huse SM, Huber JA, Welch DM, Relman DA, Sogin ML. (2008). Exploring microbial diversity and taxonomy using SSU rRNA hypervariable tag sequencing. PLoS genetics 4, e1000255.

LaMontagne MG, Michel FC, Holden PA, Reddy CA. (2002). Evaluation of extraction and purification methods for obtaining PCR-amplifiable DNA from compost for microbial community analysis. J Microbiol. Methods 49, 255-264.

Liu H, Probert I, Uitz J, Claustre H, Aris-Brosou S, Frada M, Not F, de Vargas C. (2009). Extreme diversity in noncalcifying haptophytes explains a major pigment paradox in open oceans. Proc Natl Acad Sci 106(31), 12803-12808

Liu X, Yao T, Kang S, Jiao N, Zeng Y, Liu Y. (2010). Bacterial community of the largest oligosaline lake, Namco on the Tibetan Plateau. Geomicrobiol J 27, 669- 682.

Logares, R., Shalchian-Tabrizi, K., Boltovskoy, A. & Rengefors, K. 2007. Extensive dinoflagellate phylogenies indicate infrequent marine–fresh- water transitions. Mol. Phylogenet. Evol. 45, 887–903.

Marchetto A, Mosello R, Rogora M, Manca M, Boggero A, Morabito G, Musazzi S, Tartari GA, Nocentini AM, Pugnetti A. (2004). The chemical and biological

71 response of two remote mountain lakes in the Southern Central Alps (Italy) to twenty years of changing physical and chemical climate. J Limnol 63, 77-89.

Marlowe IT, Green JC, Neal AC, Brassell SC, Eglinton G, Course PA. (1984). Long-

Chain (n-C37-C39) alkenones in the prymnesiophyceae. Distribution of alkenones and other lipids and their taxonomic significance. Br Phycol J 19, 203-216.

Moon-van der Staay S.Y., De Wachter R., Vaulot D. (2001). Oceanic 18S rDNA sequences from picoplankton reveal unsuspected eukaryotic diversity. Nature 409, 607– 610.

Müller PJ, Kirst G, Ruhland G, von Storch I, Rosell-Mele A. (1998). Calibration of k' the alkenone paleotemperature index U37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S). Geochim Cosmochim Acta 62, 1757-1772.

Newton RJ, Jones SE, Eiler A, McMahon KD, Bertilsson S. (2011). A guide to the natural history of freshwater lake bacteria. Microbiol Mol Biol R 75, 14-49.

Nold SC, Zwart G. (1998). Patterns and governing forces in aquatic microbial communities. Aquat Ecol 32, 17-35.

Not F, Latasa M, Scharek R, Viprey M, Karleskind P, et al. (2008) Phytoplankton diversity across the Indian Ocean: A focus on the picoplanktonic size fraction. Deep-Sea Res. I 55, 1456–1473.

Pawlowski J, Christen R, Lecroq B et al. (2011) Eukaryotic Richness in the Abyss, Insights from Pyrotag Sequencing. PLoS ONE 6, e18169, doi, 10.1371/journal.pone.0018169.

Prahl FG, Wakeham SG. (1987). Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330, 367-369.

Prescott, D. M. 1994. The DNA of ciliated protozoa. Microbiol. Rev. 58, 233–267.

Presthus Heggen M, Birks HH, Anderson NJ. (2010). Long-term ecosystem dynamics of a small lake and its catchment in west Greenland. Holocene 20, 1207-1222.

Pruesse E, Quast C, Knittel K, Fuchs B, Ludwig W, Peplies J, Glöckner FO. (2007). SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acid Res. 35:7188

Quayle WC, Peck LS, Peat H, Ellis-Evans JC, Harrigan PR. (2002). Extreme responses to climate change in Antarctic lakes. Science 295, 645.

72 Simon N, Campbell L, Ornolfsdottir E, Groben R, Guillou L, Lange M. and Medlin LK. (2000). Oligonucleotide probes for the identification of three algal groups by dot blot and fluorescent whole-cell hybridization. J. Eukaryot. Microbiol. 47: 76-84.

Shalchian-Tabrizi K, Reier-Røberg K, Ree DK, Klaveness D, Bråte J (2011) Marine freshwater colonizations of Haptophytes inferred from phylogeny of environmental 18S rDNA sequences. J Eukaryot. Microbiol. 58, 315-318.

Slapeta, J., Moreira, D. & Lopez-Garcia, P. 2005. The extent of protist diversity insights from molecular ecology of freshwater eukaryotes. Proc. R. Soc. Biol. Sci. Ser. B, 272, 2073–2081.

Stoeck T, Behnke A, Christen R, Amaral-Zettler L, Rodriguez-Mora M, Chistoserdov A et al. (2009). Massively parallel tag sequencing reveals the complexity of anaerobic marine protistan communities. BMC Biology 7, 72.

Stoeck T, Bass D, Nebel M, Christen R, Jones MD, Breiner HW, Richards TA (2010) Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water. Mol. Ecol. 19, 21- 31.

Straile D. (2005). Food webs in lakes—seasonal dynamics and the impact of climate variability. Aquatic food webs, an ecosystem approach. Oxford Univ. Press, Oxford 41-50.

Theroux SM, D'Andrea WJ, Toney JL, Amaral-Zettler LA, Huang Y. (2010). Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes, Implications for continental paleotemperature reconstructions. Earth Planet Sci Lett 300, 311-320.

Tyrrell, T and Merico, A. (2004) “Emiliania huxleyi: Bloom observations and the conditions that induce them” in Coccolithophores—from molecular processes to global impact. Eds. H. R. Thierstein and J. R. Young (Springer), 75–97.

Volkman JK, Eglinton G, Corner EDS, Forsberg TEV. (1980). Long-chain alkenes and alkenones in the marine coccolithophorid Emiliania-huxleyi. Phytochemistry 19, 2619-2622.

Xing P, Hahn MW, Wu QL. (2009). Low taxon richness of bacterioplankton in high- altitude lakes of the eastern Tibetan Plateau, with a predominance of Bacteroidetes and Synechococcus spp. Appl Environ Microbiol 75, 7017-7025.

Yilmaz P, Kottmann R, Field D, Knight R, Cole JR, Amaral-Zettler L, Gilbert JA, Karsch-Mizrachi I, Johnston A, Cochrane G. (2011). Minimum information about a marker gene sequence (MIMARKS) and minimum information about any (x) sequence (MIxS) specifications. Nat Biotechnol 29, 415-420.

73

Zhu F, Massana R, Not F, Marie D, Vaulot D (2005) Mapping of picoeucaryotes in marine ecosystems with quantitative PCR of the 18S rRNA gene. FEMS Microbial Ecol. 52, 79–92.

74 Table 1. BrayaSø water column C37 alkenone concentration and haptophyte rRNA gene copy number. Asterisks denote samples chosen for pyrosequencing.

[C37] Haptophyte rRNA gene Date Water depth (m) (μg/L) copies/mL 6/20/07 5 4 649 6/20/07 10* 15 1,788 6/20/07 14 4 986 6/20/07 20 1 230 6/17/09 4* 59 2,969 Isochrysis galbana control 1,493

75 Table 2. Sequencing summary, OTU distributions, and diversity estimates.

Upper and lower confidence bounds are 95%. Abbreviations for CatchAll:

Estimate = estimated total species richness; Chao1 = Chao1 estimated total species richness; ACE = ‘abundance-based coverage estimator’ estimated total species richness.

Bacteria Eukaryota

2007 2009 2007 2009 Sequenced tags 16 533 17 034 1 554 2 795 Total Observed OTUs 3 682 3 526 59 63 Singletons 2 679 2 604 18 25 Shared OTUs 799 25

Shared OTUs >1 72% 100%

Jaccard 0.125 0.258

Sorensen 0.222 0.41

Morisita-Horn 0.848

Bray-Curtis 0.527

Richness estimates 2007 2009 27 353 18 727 CatchAll Estimate (- 8148, + 12 425) (-1864, +2124) 11 142 11 060 Chao (- 838, + 945) (-866, + 979) 20 823 22 639 ACE (- 2398, + 2689) (-2749, + 3211)

76 Table 3. Most abundant bacterial and eukaryotic OTUs. Bacterial OTUs defined at 97% similarity, eukaryote OTUs defined at 96% similarity. Asterisk (*) denotes OTUs with a relative abundance difference greater than one order of magnitude. GAST = average distance between OTU and Global Alignment for

Sequence Taxonomy reference sequences (A) The 35 most abundant bacterial tags, ranked by total relative abundance. (B) The 15 most abundant eukaryote tags, ranked by total relative abundance.

A. Bacteria Taxonomy 2007 2009 Total GAST Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Sporichthyaceae 1660 1136 2796 0.0112 Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Microbacteriaceae; Limnoluna 1115 796 1911 0.0079 Bacteria; Proteobacteria; Betaproteobacteria; Burkholderiales; Burkholderiaceae; Polynucleobacter 435 660 1095 0.0054 Bacteria; Proteobacteria; Betaproteobacteria; Burkholderiales; Comamonadaceae 505 488 993 0.0066 Bacteria; Proteobacteria; Alphaproteobacteria; Rickettsiales; SAR11; Pelagibacter 550 292 842 0.0198 Bacteria; Bacteroidetes; Flavobacteria; Flavobacteriales; Cryomorphaceae; Owenweeksia 429 322 751 0.0051 Bacteria; Proteobacteria; Alphaproteobacteria; Rickettsiales; SAR11; Pelagibacter 390 308 698 0.0242 Bacteria; Proteobacteria; Betaproteobacteria; Burkholderiales; Alcaligenaceae 174 470 644 0.009 Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Sporichthyaceae 360 268 628 0.0145 Bacteria; Chloroflexi 177 434 611 0.0121 Bacteria; Bacteroidetes; Sphingobacteria; Sphingobacteriales; Cyclobacteriaceae; Algoriphagus 338 191 529 0.0114 Bacteria; Actinobacteria; Actinobacteria; Acidimicrobiales; Acidimicrobiaceae 148 317 465 0.0046 Bacteria; Bacteroidetes; Flavobacteria; Flavobacteriales; Flavobacteriaceae; Flavobacterium 420* 25 445 0.0066 Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Sporichthyaceae 215 228 443 0.0122

77 Bacteria; Proteobacteria; Gammaproteobacteria; Thiotrichales; Piscirickettsiaceae; Thiomicrospira 5 432* 437 0.0334 Bacteria; Proteobacteria; Epsilonproteobacteria; Campylobacterales; Helicobacteraceae; Sulfurovum 0 424* 424 0.0344 Bacteria; Actinobacteria; Actinobacteria; Acidimicrobiales; Iamiaceae; Iamia 254 145 399 0.0064 Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Microbacteriaceae 142 212 354 0.0086 Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Sporichthyaceae 190 110 300 0.0099 Bacteria; Actinobacteria; Actinobacteria; Acidimicrobiales; Acidimicrobiaceae 102 182 284 0.0058 Bacteria; Actinobacteria; Actinobacteria; Nitriliruptorales; Nitriliruptoraceae; Nitriliruptor 74 171 245 0.0174 Bacteria; Proteobacteria; Epsilonproteobacteria; Campylobacterales; Helicobacteraceae; Sulfuricurvum 0 237* 237 0.0184 Bacteria; Firmicutes; Erysipelotrichi; Erysipelotrichales; Erysipelotrichaceae 90 135 225 0.1072 Bacteria; Bacteroidetes; Flavobacteria; Flavobacteriales; Flavobacteriaceae; Lutibacter 56 162 218 0.0102 Bacteria; Bacteroidetes; Sphingobacteria; Sphingobacteriales; Cytophagaceae; Adhaeribacter 116 99 215 0.0389 Bacteria; Bacteroidetes; Sphingobacteria; Sphingobacteriales 127 87 214 0.0219 Bacteria; Bacteroidetes; Sphingobacteria; Sphingobacteriales; Sphingobacteriaceae; Sphingobacteriaceae; Pedobacter 146 68 214 0.005 Bacteria; Firmicutes; Erysipelotrichi; Erysipelotrichales; Erysipelotrichaceae 99 107 206 0.0422 Bacteria; Proteobacteria; Betaproteobacteria; Methylophilales; Methylophilaceae 52 149 201 0.0074 Bacteria; Bacteroidetes; Flavobacteria; Flavobacteriales; Flavobacteriaceae; Flavobacterium 49 129 178 0.0267 Bacteria; Planctomycetes; Planctomycetacia; Planctomycetales; Planctomycetaceae; Gemmata 9 163* 172 0.0058 Bacteria; Verrucomicrobia; Verrucomicrobiae; Verrucomicrobiales; Verrucomicrobiaceae 9 159* 168 0.0133 Bacteria; Actinobacteria; Actinobacteria; Actinomycetales; Sporichthyaceae 97 66 163 0.0421 Bacteria; Proteobacteria; Betaproteobacteria; Burkholderiales; Burkholderiaceae; Polynucleobacter; cosmopolitanus 144* 14 158 0.0159 Bacteria;Planctomycetes;Planctomycetacia; Planctomycetales;Planctomycetaceae;Pirellula 71 70 141 0.0273

78 B. Eukaryote Taxonomy 2007 2009 Total GAST Eukaryota; stramenopiles; Bacillariophyta; Fragilariophyceae 1518 929 2447 0.0059 Eukaryota; Alveolata; Ciliophora 864 597 1461 0.0577 Eukaryota; stramenopile 640 262 902 0.1185 Eukaryota; Metazoa; Arthropoda 0 897* 897 0.0189 Eukaryota; Viridiplantae; Chlorophyta; Chlorophyceae; Chlamydomonadales 10 501* 511 0.01 Eukaryota; Alveolata; Ciliophora 35 185 220 0.0189 Eukaryota; Alveolata; Ciliophora 1 181* 182 0.0141 Eukaryota; Haptophyceae 155 16 171 0.0179 Eukaryota; Katablepharidophyta; Katablepharidaceae 120 41 161 0.1036 Eukaryota; Viridiplantae; Chlorophyta; Chlorophyceae; Chlamydomonadales 33 110 143 0.0071 Eukaryota; Alveolata; Ciliophora 70 60 130 0.1668 Eukaryota; Cryptophyta 99* 8 107 0.0107 Eukaryota; Alveolata 1 65* 66 0.0856 Eukaryota; stramenopiles; Chrysophyceae 0 47* 47 0.0539 Eukaryota; stramenopiles; Chrysophyceae 29 16 45 0.0287

79

Figure 1. Site map showing Lake BrayaSø in the Kangerlussuaq region,

Greenland. Modified from D’Andrea et al., 2006.

80

June 20, 2007 ! June 17, 2009! 0! 10! 20! 30! 40! 0! 10! 20! 30! 40! 50! 60! 70! 0! 0! !

5! ! 5! Temp (°C)! 10! 10! pH! Depth (m) Depth

Depth (m) Depth Chl ("g/L)! Conductivity (mS/cm)! 15! 15! DO (%*0.1)! C37 alk ("g/L)! 20! 20! rRNA gene copies (*0.01)!

Figure 2. BrayaSø water column. Left panel shows 2007 sampling profile, and right panel shows 2009 sampling profile. Arrows denote sample depth for the samples we sequenced Dissolved oxygen (DO) is decreased one order of magnitude and haptophyte rRNA gene copy number is decreased two orders of magnitude to plot along the same axis.

81

Figure 3. (A) June average air temperatures for 2007 and 2009. (B) Average monthly air temperatures for 2007 and 2009. Station data from Kangerlussuaq,

Greenland (67.017° N, -50.700° W). http://www.ncdc.noaa.gov/oa/ncdc.html.

Arrows denote sampling dates.

82

Figure 4. Venn diagrams of microbial communities. (A) Overlap between 2007 and 2009 bacterial communities, OTUs defined at 97% similarity (B) As in (A) with singletons removed (C) Overlap between 2007 and 2009 eukaryotic communities, OTUs defined at 94% similarity, (D) as in (C) with singletons removed.

83

CHAPTER 4

PRODUCTION AND TEMPERATURE SENSITIVITY OF LONG CHAIN

ALKENONES IN CULTURES OF PSEUDOISOCHRYSIS PARADOXA, NOMEN

NUDUM

SUSANNA THEROUX1,2

JAIME TONEY1,‡

LINDA AMARAL-ZETTLER1,2,3

YONGSONG HUANG1

1 BROWN UNIVERSITY, Department of Geological Sciences

2 MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

3 BROWN UNIVERSITY, Department of Ecology and Evolutionary Biology

‡ Present address: UNIVERSITY OF GLASGOW, School of Geographical and Earth Sciences

Prepared for:

84 Organic Geochemistry

85 Abstract

k k' The alkenone unsaturation index (U37 and U37 ) serves as a critical tool for reconstructing paleotemperatures in marine environments. Lacustrine haptophyte species are distinct from their ubiquitous and well-studied marine counterparts, and the unknown genotypic effects on long chain alkenone production by lacustrine

k species have hindered the application of the U37 paleotemperature proxy to lake sediment records. Previous work in Lake George, North Dakota revealed the presence of two novel alkenone-producing haptophyte species, one that is closely related to the cultured organism Pseudoisochrysis paradoxa CCMP715. Like its close relative

Isochrysis galbana, P. paradoxa is an alkenone producer, with a lipid fingerprint abundant in tetraunsaturated alkenones. We present here the first calibration of the

k U37 paleotemperature proxy for a culture of P. paradoxa. While the extent of its geographic distribution remains unclear, P. paradoxa undoubtedly contributes to sedimentary archives throughout the world’s lacustrine regions, thus confirming the

k importance of culture-based and in situ U37 calibrations in lake environments.

86 1. Introduction

For over two decades, alkenone lipids have enabled the reconstruction of sea surface temperatures from marine sediments. Haptophyte algae in the group

Isochrysidales are the exclusive producers of alkenone lipids, an extremely specific biomarker found throughout the world’s pelagic, coastal, and lacustrine

k environments. The paleotemperature proxy U37 incorporates the relative abundance

of the di-(C37:2), tri-(C37:3) and tetra-(C37:4) unsaturated alkenones as a proxy for water temperature (Brassell et al., 1986), with greater degrees of unsaturation occurring at

k' lower temperatures (Marlowe, 1984; Brassell et al., 1986). A modified U37 (Prahl and

Wakeham, 1987; Prahl et al., 1988) has been widely applied to marine sediments where the tetraunsaturated alkenone is largely absent. The marine calibration of the

k' U37 is used universally (Prahl and Wakeham, 1987; Muller et al., 1988), the consequence of cosmopolitan Emiliania huxleyi and Gephyrocapsa oceanica being responsible for the majority of alkenone production in marine waters.

The recent discovery of widespread alkenone lipids distributions in lake sediments (Cranwell, 1985; Zink et al., 2001; Chu et al., 2005; D’Andrea and Huang,

2005; Pearson et al., 2008; Toney et al., 2010; Theroux et al., 2010; Toney et al.

k 2011) has generated interest in the applicability of the U37 proxy to lake sediment archives. Alkenone lipids in lakes have been shown to reflect mean annual air temperature (Chu et al., 2005) and in situ lake water temperature (Toney et al., 2010;

k D’Andrea et al., 2011). However, there is no universal U37 calibration for all lake environments, largely the result of the genetic diversity of the lacustrine haptophytes

87 (Theroux et al., 2010). The abundance of novel haptophyte species in these lake

k environments creates a new demand for in situ and culture-based U37 calibrations.

A genetic survey of Lake George, ND surface sediments revealed the presence of two haptophyte algal species responsible for alkenone production, one that is closely related (99% similar over 461bp of 18S rRNA gene sequence) to

Pseudoisochrysis paradoxa (Theroux et al., 2010). Originally isolated from the brackish York River Estuary in the Chesapeake Bay, Virginia, USA, P. paradoxa has never been formally described (nomen nudum; Jordan et al., 2004). The 18S ribosomal RNA (rRNA) gene sequence for P. paradoxa CCMP715 (Genbank

AM490999; Medlin et al., 2008) is 100% identical to that of the coastal/lacustrine haptophyte Isochrysis galbana (Genbank HM149541), highlighting the difficulty in distinguishing haptophyte species through 18S rRNA gene sequences alone.

Isochrysis galbana is a known alkenone producer, and P. paradoxa was reported to

k have alkenones (Marlowe et al., 1984), although its U37 temperature calibration has never been determined.

The global extent of P. paradoxa populations is currently unknown; if it ecologically resembles its close relative, I. galbana, it can survive in a wide range of brackish environments and marine environments and contribute to the alkenone sediment records (Marlowe et al., 1990; Versteegh et al., 2001; Liu et al., 2009).

k Only a few species of haptophytes have been grown in culture to determine their U37 calibration: ubiquitous marine species Emiliania huxleyi (Prahl et al., 1988; Prahl et al., 2003; Conte et al., 1998) and Gephyrocapsa oceanica (Sawada et al., 1996;

Conte et al., 1998), and lacustrine/brackish species Isochrysis galbana (Versteegh et

88 al., 2001) and Chrysotila lamellosa (Sun et al., 2007). Paleoclimate work in these brackish environments requires a better understanding of the ecology and controls on alkenone unsaturation of these haptophyte species. To this end, we grew P. paradoxa at a variety of temperatures to determine its alkenone unsaturation-temperature relationship.

2. Methods

2.1 Pseudoisochrysis paradoxa cultures

We purchased cultures of Pseudoisochrysis paradoxa (CCMP715, also known as

CCAP 949/1, CCAPVA12, UTEX 1988) from the Provasoli-Guillard National Center for Marine Algae and Microbiota. We grew cultures in 0.2 µm filter-sterilized seawater amended with f/2 nutrients (Guillard, 1975) in full spectrum light on a 24:0 light:dark cycle. We verified the media was alkenone-free. We grew these batch cultures in triplicate volumes of 50 ml at 5, 10, 15, 21 and 24°C. We initiated cultures at equal cell concentrations of 8000 cells/ml using an inoculum from a culture previously acclimated to a given temperature. We monitored cell concentrations for three weeks using haemocytometer counts to ensure the cultures remained in the exponential growth phase. At the end of three weeks we harvested the cultures for alkenone analysis. One of the three 5°C cultures was discarded because it failed to grow.

2.2 Lipid analysis of cultures

89 We filtered culture material onto precombusted 47 mm glass microfiber filters glass fiber filters (Whatman, Piscataway, NJ) immediately froze at -20°C and then freeze- dried the filters overnight (Labconco, Kansas City, MO). We extracted the filters using three 20-minute bursts of sonication in dichloromethane (DCM), and ran the total lipid extracts on an Agilent 6890plus Gas Chromatograph Flame Ionization

Detector (GC-FID) for detection and quantification of alkenones using an internal C36

k n-alkane standard and an external alkenone standard of known U37 -temperature value to ensure analytical precision (< 0.1°C). We used a Varian VF200 60m GC column

(60m × 250μm × 0.10 µm) with the following temperature program: an initial temperature of 100°C (hold 1-min), ramp 20°C/min to 200°C (hold 1-min), ramp

4°C/min to 320°C (hold 0 min), hold at 320°C for 5-min.

3. Results

3.1 Cultures

Triplicate cultures of P. paradoxa behaved similarly within temperature regimes

(Table 1). Growth rate and alkenone concentrations per cell displayed an inverse relationship (Figure 1), with growth rate highest at 21°C (54.30 divisions/day) and lowest at 5°C (0.01 div/day) and alkenone per cell concentrations were highest at 5°C

(1.961 pg/cell) and lowest at 21°C (0.042 pg/cell). Final cell counts were highest at

21°C and lowest at 5°C (Table 1).

k k' 3.2 U37 and U37 calibrations

90 k k' We calculated the U37 and U37 calibrations for the P. paradoxa cultures and plotted

k k these versus growth temperature (Figure 2). The U37 calibration (U37 = 0.0226T-

2 k' 0.5149, R = 0.905, RMSE = 0.05) was slightly more robust than the U37 calibration

k' 2 2 k (U37 = 0.001T -0.0256T+0.1754, R = 0.89, RMSE = 0.02), although only the U37

temperature calibration fit a linear regression. The C37:4 alkenone comprised almost

40% of the C37 alkenones in the 5°C culture (Figure 3), although at all temperatures

k the C37:3 was the dominant alkenone homolog (Figure 4). The U37 calibration for P. paradoxa had a similar slope to the in situ calibration from Lake George, ND (Figure

5; Toney et al., 2010), and clustered with other lake-based calibrations that were all

k' distinct from the marine calibration. The U37 calibration was also distinct from

k' previously reported U37 calibrations from other haptophyte cultures and environmental samples (Figure 6), and was best fit with a second order polynomial

k regression (Figure 2) whose y-intercept was closest to the marine U37 calibration.

4. Discussion

4.1 Growth stage and alkenone production

The P. paradoxa cultures demonstrated an enhanced alkenone production per cell at the slowest growth rates in the 5°C culture. This agrees with previous observations in cultures of E. huxleyi and G. oceanica that demonstrated highest alkenone concentrations per cell with lowest growth temperatures (Conte et al.,

1998). However, low growth temperature often corresponds to low growth rates, so it is unclear if low growth temperature alone will result in enhanced alkenone

91 accumulation. Alkenones are presumed to serve as an energy storage molecule in haptophytes (Epstein et al., 2001; Eltgroth et al., 2005) and cellular production of alkenones increases during stationary growth phase and decreases after cultures are placed in the dark (Epstein et al., 2001; Eltgroth et al., 2005). The accumulation of alkenones at low temperatures and low growth rates may be the result of photosynthetic energy input exceeding the cell’s capacity for growth and division

(Roessler, 1990) as was observed at both low temperatures and low nutrient

k 2 conditions. The P. paradoxa U37 calibration R value of 0.905 supports the robustness of the alkenone temperature-dependent relationship. The coherency in P. paradoxa biological replicates further supports this linear relationship.

k k' The affect of growth rate on haptophyte U37 and U37 is unclear (Conte et al.,

1995; Epstein et al., 1998; Popp et al., 1998). In this study, we used a 24:0 light to dark regime to eliminate alkenone metabolism during darkness. In batch culture

k' experiments, U37 decreases under nutrient stress and increases under prolonged darkness (Prahl et al., 2006), both conditions that may result in slower growth rates.

Cultures of various strains of E. huxleyi grown in 12:12 or 0:24 light to dark regimes

k' exhibited contrasting increases or decreases in U37 values depending on light regime

k' (Epstein et al., 2001). These fluctuations in U37 were 0.013 to 0.029 units, a small amount, and therefore we do not believe the 24-hour light regime exerted a significant

k change in U37 values. Previous studies have shown that differences can exist between batch and continuous-culture methods (Popp et al., 1998) although it is debated which of these two methods more accurately replicates the natural environment. Growth

92 phase also has been shown to influence alkenone unsaturation in batch culture (Conte et al., 1998; Epstein et al., 1998), although continuous culture, and therefore constant

k growth state, imparted no change on U37 values (Popp et al., 1998). To control for this variation we harvested all cultures while in exponential growth phase.

4.2 Comparison to other species

Pseudoisochrysis paradoxa resembled other lacustrine haptophytes in the

presence of the C37:4 alkenone (Cranwell, 1985; Li et al., 1996; Zink et al., 2001).

Similar to its close relative I. galbana, P. paradoxa cultures had predominant C37:3

alkenones and the absence of the C38 methyl ketone (Figure 3). The abundance of the

k C37:4 alkenone in the P. paradoxa cultures may explain the strong U37 calibration and

k' k weaker U37 calibration (Figure 2). The P. paradoxa U37 calibration matched

k previously reported lacustrine U37 calibrations (Figure 5), and had the same slope as the in situ calibration generated in Lake George, ND (Toney et al., 2010). This result

was surprising as the Lake George in situ calibration was dominated by C37:4 alkenones, whereas P. paradoxa cultures had abundant but not dominant C37:4 alkenones. Their similarity in slope suggests the importance of incorporating the C37:4 homolog into the calibration equation when its abundance is high (>20% of total C37

k' alkenones). The U37 calibration was best fit with a second-order polynomial regression, indicating a non-linear relationship between temperature and alkenone unsaturation. At lower temperatures, the P. paradoxa cultures actually had decreasing

k' U37 values with increasing temperatures. This result indicates that the inclusion of the

93 tetraunsaturated alkenone is critical to the accurate reconstruction of temperature in cultures of P. paradoxa.

While P. paradoxa cultures had abundant C37:4 alkenones, sediments and water samples from Lake George, ND (Toney et al., 2010), as well as Lake BrayaSø in Greenland (D’Andrea and Huang, 2005; D’Andrea et al., 2011) and Ace Lake,

Antarctica (Coolen et al., 2004) contain alkenone signatures with dominant C37:4.

Previously, percent C37:4 (C37:4/C37:4+C37:3+C37:2) was proposed as a paleosalinity proxy

(Schulz et al., 2000; Roselle-Mele et al., 2002) although the lack of correlation

between salinity and percent C37:4 in a global array of lake systems (Mercer et al.,

2005; Chu et al., 2005; Toney et al., 2010; Theroux et al., 2010) suggests that the

relationship between C37:4 and salinity is a result of alkenone-producing species distributions across salinity tolerances. Percent C37:4 can range from 8-53% in C. lamellosa (Rontani et al., 2004) and 0-34% in I. galbana (Marlowe et al., 1984), and

up to 96% in a series of lakes in China (Chu et al., 2005). In this study, percent C37:4 ranged from 6% at 24°C to 40% at 5°C (Table 1, Figure 4). The alkenone concentrations per cell were within the range observed for I. galbana (0.0098-0.61 pg/cell; Versteegh et al., 2001). The highest alkenone/cell concentrations observed for P. paradoxa (1.961 pg/cell) was close to the range observed in an Isochrysis sp. culture (1.8 pg/cell; Marlowe, 1984).

4.3 Application to the natural environment

The similarity of the P. paradoxa calibration to the in situ Lake George calibration further confirms the utility of in situ calibrations. More than one alkenone

94 producer occurs in Lake George waters (Theroux et al., 2010; Toney et al., 2012), and the dual methods of alkenone production by two distinct species likely explains

k the offset between the Lake George in situ U37 calibration and the P. paradoxa culture-based calibration. The Lake George alkenone fingerprint was dominated by

the tetraunsaturated C37:4 alkenone, unlike P. paradoxa cultures with abundant C37:4 but dominant C37:3 (Toney et al., 2010; Theroux et al. 2010; Toney et al., 2011). The presence of the C37:4 homolog in both Lake George waters and the P. paradoxa

k k' cultures supports the use of the U37 proxy instead of the U37 index. Of note, alkenone production in Lake George occurs during a brief window of time following spring ice-off (Toney et al., 2010), and alkenone records likely reflect spring bloom water temperatures alone.

The P. paradoxa alkenone “fingerprint” closely resembled the alkenone

k' signature of I. galbana, but their U37 calibrations are very different (Figure 5).

Studies have shown that different geographically isolated strains of the same species of haptophyte, E. huxleyi and G. oceanica, have different modes of alkenone unsaturation with temperature (Conte et al., 1995; Conte et al., 2006). Therefore, it is no surprise that different but closely related haptophytes will also possess different

k k U37 calibrations. The slopes of many U37 calibrations across haptophyte species are similar (Figure 5). This suggests that the temperature response of haptophytes is

k consistent, but the determining factor of the U37 calibration y-intercept, is still

k unknown. Reconstructing temperature from organisms with similar U37 calibration

95 slopes will prevent the reconstruction of absolute temperature but may still be able to reconstruct relative temperature change.

5. Conclusions

k The application of the U37 paleotemperature proxy to lake sediments depends

k upon a robust U37 calibration. As evidenced by studies to date, lakes are not uniform in their alkenone-producing haptophyte populations, thereby complicating the use of

k a universal lake U37 proxy calibration. Our study is one of few to cultivate an

k individual haptophyte species at various temperatures to calibrate the U37 - temperature relationship. Pseudoisochrysis paradoxa has proven to be distinct in its mode of alkenone unsaturation, with a calibration equation that is significantly different from its close genetic relative Isochrysis galbana. The similarity between the Lake George in situ calibration and the P. paradoxa calibration reported in this

k study gives credence to the applicability of in situ U37 calibrations. As with all alkenone-based temperature reconstructions, it is important to verify the continuity of alkenone distributions throughout downcore sediments in order to apply a single calibration to the temperature reconstruction. We anticipate future studies comparing

k both in situ and culture-based U37 calibrations to further resolve species-specific modes of alkenone production in lake environments.

Acknowledgements

This work was supported by a National Science Foundation award to Y. Huang

(EAR-1122749) and L. Amaral-Zettler (EAR-1124192), a Brown SEED fund to Y.

96 Huang and L. Amaral-Zettler, and an American Association of University Women dissertation fellowship to S. Theroux.

References

Brassell SC, Eglinton G. (1986). Molecular geochemical indicators in sediments. Acs Symposium Series 305:10-32.

Brassell SC, Eglinton G, Marlowe IT, Pflaumann U, Sarnthein M. (1986). Molecular stratigraphy - a new tool for climatic assessment. Nature 320:129-133.

Chu GQ, Sun Q, Li SQ, Zheng MP, Jia XX, Lu CF, Liu JQ, Liu TS. (2005). Long- chain alkenone distributions and temperature dependence in lacustrine surface sediments from China. Geochimica Et Cosmochimica Acta 69:4985-5003.

Conte MH, Sicre MA, Ruhlemann C, Weber JC, Schulte S, Schulz-Bull D, Blanz T. k' (2006). Global temperature calibration of the alkenone unsaturation index U37 in surface waters and comparison with surface sediments. Geochemistry Geophysics Geosystems 7:doi:10.1029/2005GC001054.

Conte MH, Thompson A, Lesley D, Harris RP. (1998). Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica. Geochimica Et Cosmochimica Acta 62:51-68.

Conte MH, Thompson A, Eglinton G, Green JC. (1995). Lipid biomarker diversity in the coccolithophorid Emiliania huxleyi (Prymnesiophyceae) and the related species Gephyrocapsa oceanica.Journal of Phycology 31:272–282.

Coolen MJL, Muyzer G, Rijpstra WIC, Schouten S, Volkman JK, Sinninghe Damsté JS. (2004). Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth and Planetary Science Letters 223:225-239.

Cranwell PA. (1985). Long-Chain Unsaturated-Ketones in Recent Lacustrine Sediments. Geochimica Et Cosmochimica Acta 49:1545-1551.

D'Andrea WJ, Huang Y. (2005). Long chain alkenones in Greenland lake sediments: Low δ13C values and exceptional abundance. Organic Geochemistry 36:1234- 1241.

D'Andrea WJ, Huang Y, Fritz SC, Anderson NJ. (2011). Abrupt Holocene climate

97 change as an important factor for human migration in West Greenland. Proceedings of the National Academy of Sciences

D’Andrea WJ, Huang Y, Fritz SC, Anderson NJ. (2011). Abrupt Holocene climate change as an important factor for human migration in West Greenland. Proceedings of the National Academy of Sciences 108:9765.

Eltgroth ML, Watwood RL, Wolfe GV. (2005). Production and cellular localization of neutral long-chain lipids in the haptophyte algae Isochrysis galbana and Emiliania huxleyi. Journal of Phycology 41:1000-1009.

Epstein BL, D'Hondt S, Hargraves PE. (2001). The possible metabolic role of C-37 alkenones in Emiliania huxleyi. Organic Geochemistry 32:867-875.

Epstein BL, D'Hondt S, Quinn JG, Zhang JP, Hargraves PE. (1998). An effect of dissolved nutrient concentrations on alkenone-based temperature estimates. Paleoceanography 13:122-126.

Guillard RRL (1975). Culture of phytoplankton for feeding marine invertebrates. In Culture of marine invertebrate animals. WL Smith, MH Chanley, eds. Plenum Press, New York, p. 26.

Jordan RW, Cros L, Young JR. (2004). A revised classification scheme for living haptophytes. Micropaleontology 50:55-79.

Li JG, Philp RP, Pu F, Allen J. (1996). Long-chain alkenones in Qinghai Lake sediments. Geochimica Et Cosmochimica Acta 60:235-241.

Liu Y, Yao T, Zhu L, Jiao N, Liu X, Zeng Y, Jiang H. (2009). Bacterial diversity of freshwater alpine lake Puma Yumco on the Tibetan Plateau. Geomicrobiology Journal 26:131-145.

Marlowe IT, Brassell SC, Eglinton G, Green JC. (1990). Long-chain alkenones and alkyl alkenoates and the fossil coccolith record of marine sediments. Chemical Geology 88:349 - 375.

Marlowe IT, Green JC, Neal AC, Brassell SC, Eglinton G, Course PA. (1984). Long- Chain (n-C37-C39) alkenones in the prymnesiophyceae. Distribution of alkenones and other lipids and their taxonomic significance. British Phycological Journal 19:203-216.

Medlin LK, Saez AG, Young JR. (2008). A molecular clock for coccolithophores and implications for selectivity of phytoplankton extinctions across the K/T boundary. Marine Micropaleontology 67:69-86.

Mercer JL, Zhao MX, Colman SM. (2005). Seasonal variations of alkenones and U-

98 37(K) in the Chesapeake Bay water column. Estuarine Coastal and Shelf Science 63:675-682.

Müller PJ, Kirst G, Ruhland G, von Storch I, Rosell-Mele A. (1998). Calibration of k' the alkenone paleotemperature index U37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S). Geochimica Et Cosmochimica Acta 62:1757-1772.

Pearson EJ, Juggins S, Farrimond P. (2008). Distribution and significance of long- chain alkenones as salinity and temperature indicators in Spanish saline lake sediments. Geochimica Et Cosmochimica Acta 72:4035-4046.

Popp BN, Kenig F, Wakeham SG, Laws EA, Bidigare RR. (1998). Does Growth Rate Affect Ketone Unsaturation and Intracellular Carbon Isotopic Variability in Emiliania Huxleyi? Paleoceanography 13:35-41.

Prahl FG, Wolfe GV, Sparrow MA. (2003). Physiological impacts on alkenone paleothermometry. Paleoceanography 18:doi:10.1029/2002PA000803.

Prahl FG, Muehlhausen LA, Zahnle DL. (1988). Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochimica Et Cosmochimica Acta 52:2303-2310.

Prahl FG, Wakeham SG. (1987). Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330:367-369.

Prahl FG, Mix AC, Sparrow MA. (2006). Alkenone paleothermometry: Biological lessons from marine sediment records off western South America. Geochimica et Cosmochimica Acta 70:101-117.

Roessler PG. (1990). Environmental control of glycerolipid metabolism in microalgae: commercial implications and future research directions. Journal of Phycology 26:393–399.

Rontani JF, Beker B, Volkman JK. (2004). Long-chain alkenones and related compounds in the benthic haptophyte Chrysotila lamellosa Anand HAP 17. Phytochemistry 65:117-126.

Rosell-Melé A, Jansen E, Weinelt M. (2002). Appraisal of a molecular approach to infer variations in surface ocean freshwater inputs into the North Atlantic during the last glacial. Global and Planetary Change 34:143-152.

Sawada K, Handa N, Shiraiwa Y, Danbara A, Montani S. (1996). Long-chain alkenones and alkyl alkenoates in the coastal and pelagic sediments of the northwest North Pacific, with special reference to the reconstruction of Emiliania huxleyi and Gephyrocapsa oceanica ratios. Organic Geochemistry

99 24:751 - 764.

Schulz H-M, Schöner A, Emeis K-C. (2000). Long-chain alkenone patterns in the Baltic sea--an ocean-freshwater transition. Geochimica Et Cosmochimica Acta 64:469-477.

Sun Q, Chu GQ, Liu GX, Li S, Wang XH. (2007). Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellosa (Haptophyceae). Organic Geochemistry 38:1226-1234.

Theroux SM, D'Andrea WJ, Toney JL, Amaral-Zettler LA, Huang Y. (2010). Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes: Implications for continental paleotemperature reconstructions. Earth and Planetary Science Letters 300:311-320.

Toney JL, Huang Y, Fritz SC, Baker PA, Grimm E, Nyren P. (2010). Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes in the interior United States. Geochimica Et Cosmochimica Acta

Toney JL, Theroux S, Andersen RA, Coleman A, Amaral-Zettle L, Huang Y. (2012). Culturing of the first 37:4 predominant lacustrine haptophyte: Geochemical, biochemical, and genetic implications. Geochimica et Cosmochimica Acta 78:51 - 64.

k' Versteegh GJM, Riegman R, de Leeuw JW, Jansen JHF. (2001). U37 values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations. Organic Geochemistry 32:785-794.

Zink KG, Leythaeuser D, Melkonian M, Schwark L. (2001). Temperature dependency of long-chain alkenone distributions in recent to fossil limnic sediments and in lake waters. Geochimica Et Cosmochimica Acta 65:253-265.

100 Table 1. Average growth rate, cell concentrations and alkenone concentration of

Pseudoisochrysis paradoxa. SD 2% 1% 0.4% 0.4% 0.3% 6% 39% 31% 25% 13% % C37:4 SD 0.035 0.004 0.004 0.014 0.019 0.088 Uk37 -0.350 -0.290 -0.237 -0.075 SD 0.601036 0.003440 0.018491 0.010556 0.030216 1.961 0.092 0.112 0.042 0.202 (pg/cell) Alkenones cultures. All cultures were grown in triplicate. Standard Standard growntriplicate. were in cultures All cultures. 6 18 39 74 SD 109 P. paradoxa P. 20 373 960 406 532 (g/L) Alkenones SD 0.00 92,376 871,780 461,880 2,000,000 10,000 (cells/ml) 4,053,333 8,600,000 2,666,667 10,000,000 Final cell concentration concentration cell Final SD 0.00 0.50 4.74 2.51 10.87 0.01 21.99 46.70 54.30 14.45 (div/day) Growth rate rate Growth 5 10 15 21 24 (°C) Temperature Temperature deviation (SD) as noted. deviation Table 1. Average growth rate, cell concentrations, and alkenone concentrations for concentrations alkenone and concentrations, cell growth rate, Average 1. Table

101

Figure 1. Alkenone cell concentration versus growth rate in P. paradoxa cultures.

102

k Figure 2. (Left panel) Pseudoisochrysis paradoxa U37 calibration (Right panel)

k' Pseudoisochrysis paradoxa U37 calibration.

103

C37:3

C37:4 5μm C38:3et Relative abundance

C38:4

C38:2 C37:2

Retention time

Figure 3. Gas chromatogram of P. paradoxa cultures at 10°C. Inset:

Photomicrograph of P. paradoxa culture. Scale bar 5µm.

104

Figure 4. Average alkenone concentrations for P. paradoxa cultures at different temperatures.

105 1! German lakes! Lake George! 0.8! Lake BrayaSø! 0.6! E. huxleyi! P. paradoxa! 0.4! Polar waters! C. lamellosa! 0.2!

! 0!

Uk37 -0.2!

-0.4!

-0.6!

-0.8!

-1!

-1.2! 0! 5! 10! 15! 20! 25! Temperature (°C)!

k Figure 5. Comparison of P. paradoxa U37 calibration with other species and lake- based calibrations. References are as follows: German Lakes, Zink et al., 2001;

Lake George in situ, Toney et al., 2010; Lake BrayaSø, D'Andrea et al., 2011; E. huxleyi, Prahl et al., 1988; P. paradoxa, this paper; Polar waters, Sikes and

Volkman 1993; Lake George in situ 2011, Theroux, this work; C. lamellosa, Sun et al., 2007.

106

1.2! P. paradoxa! German lakes! 1! Chinese lakes! Chinese freshwater/brackish! Chinese saline! 0.8! C. lamellosa! Global marine! 0.6! I. galbana! G. oceanica! ! 0.4! Uk'37 0.2!

0!

-0.2!

-0.4! 0! 5! 10! 15! 20! 25!

Temperature (°C)!

k' Figure 6. Comparison of P. paradoxa U37 calibration with other species and lake- based calibrations. References are as follows: P. paradoxa, this paper; German lakes, Zink et al., 2001;Chinese lakes, Chu et al., 2005; Chinese freshwater/brackish lakes, Chu et al., 2005; Chinese saline, Chu et al., 2005; C. lamellosa, Sun et al., 2007;

Global marine, Prahl and Wakeham 1987; I. galbana; Versteegh et al., 2001; G. oceanica, Sawada et al., 1996.

107 CHAPTER 5

HAPTOPHYTE DIVERSITY ALONG A SALINITY GRADIENT IN LAKES OF THE

QINGHAI REGION

SUSANNA THEROUX1,2

ELIZABETH K. THOMAS1

LINDA AMARAL-ZETTLER1,2,3

YONGSONG HUANG1

1 BROWN UNIVERSITY, Department of Geological Sciences

2 MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

3 BROWN UNIVERSITY, Department of Ecology and Evolutionary Biology

Prepared for: Geochimica et Cosmochimica Acta

108 Abstract

k k' The alkenone unsaturation index (U37 or U37 ) serves as a robust paleotemperature proxy

in aquatic environments. In environments characterized by high C37:4 alkenone concentrations, a C37:4-based paleosalinity proxy has been proposed as a means to reconstruct ancient salinity levels from lake sediments. The Qinghai region on the

Chinese-Tibetan plateau provides an ideal environment to test this proposed paleosalinity proxy given its abundance of alkenone-containing lakes, its range of lake salinities and a shared mean annual air temperature among the lakes. We sampled Lake Qinghai and three of its satellite lakes to gauge alkenone-producing haptophyte species across a salinity gradient. Our geochemical results agreed with previous research that

demonstrated a relationship between C37:4 alkenone concentrations and water salinity.

However, our DNA sequencing analyses revealed the presence of multiple alkenone- producing haptophytes, with different alkenone profiles, in a single lake. Instead of a

single haptophyte producing varying alkenone signatures, we propose that C37:4 alkenones trend with lake salinity as a result of shifts in haptophyte community composition. This study is the first to identify the haptophyte species responsible for alkenone production in

Lake Qinghai and provides an alternate explanation to the observed relationship between

lacustrine C37:4 concentrations and salinity.

109 1. Introduction

For over two decades, alkenone lipids have enabled the reconstruction of sea surface temperatures from marine sediments. Haptophyte algae in the group

Isochrysidales are the exclusive producers of alkenone lipids, an extremely specific biomarker found throughout the world’s pelagic, coastal, and lacustrine environments.

k The paleotemperature proxy U37 incorporates the relative abundance of the di-(C37:2), tri-

(C37:3) and tetra-(C37:4) unsaturated alkenones as a proxy for water temperature (Brassell et al., 1986), with greater degrees of unsaturation occurring at lower temperatures

k' (Marlowe, 1984; Brassell et al., 1986). A modified U37 (Prahl and Wakeham, 1987; Prahl et al., 1988) has been widely applied to marine sediments where the tetraunsaturated

k' alkenone is largely absent. The marine calibration of the U37 is used universally (Prahl and Wakeham, 1987; Muller et al., 1988), the consequence of cosmopolitan Emiliania huxleyi and Gephyrocapsa oceanica being responsible for the majority of alkenone production in marine waters. In lacustrine waters, Isochrysis galbana and Chrysotila lamellosa (Marlowe et al., 1984) are known alkenone-producers, and recent research has identified novel species of haptophytes that fall within the Isochrysidales and are therefore potential alkenone producers (Coolen et al., 2004; D’Andrea et al., 2006;

Theroux et al., 2010).

Dominant C37:4 alkenone abundances (%37:4 > (C37:3+C37:2)) have been reported in many lacustrine settings (Cranwell, 1985; Volkman et al., 1988; Li et al.,1996; Thiel et al., 1997; Wang and Zheng, 1998; Sun et al., 2004; Coolen et al., 2004; Chu et al, 2005;

D’Andrea and Huang 2005; Liu et al., 2006; Toney et al., 2010). In contrast, abundant or

dominant C37:4 concentrations in marine environments are restricted to the North Atlantic

110 and Nordic Seas (Rosell-Mele, 1998; Sicre et al., 2002) and nearshore environments

(Schulz et al., 2000; Chu et al., 2005; Mercer et al., 2005). The accumulation of more unsaturated alkenones has previously been observed when haptophytes are grown under nutrient deplete conditions, and therefore lower growth rates (Prahl et al., 2003; Prahl et

al., 2006). However, given their abundance in low salinity waters, C37:4 alkenones have been proposed as a paleosalinity proxy, with C37:4 concentrations decreasing as salinity increases. Since first proposed by Rosell-Mele et al., (2002), numerous studies have attempted to confirm or refute this salinity proxy and the results are still inconclusive. On

regional scales, the %C37:4 ([C37:4] / [C37:2]+[C37:3]+[C37:4]) salinity relationship has been observed in northern high latitude oceans (Rosell-Mele, 1998; Sicre et al., 2002; Harada et al., 2003) and has been used to reconstruct paleo-salinity in the Sea of Okhotsk (Seki et

al., 2005). Concentrations of C37:4 fail to accurately reconstruct salinity levels from lakes in Greenland (D’Andrea and Huang 2005) and Chesapeake Bay (Mercer et al., 2005) and

a linear relationship between salinity and %C37:4 was not found in a global survey of alkenone-containing lake sediments (Theroux et al., 2010) and a survey of lakes in the

Northern Great Plains (Toney et al., 2010).

Temperature effects may overwrite patterns in C37:4 concentrations with salinity

(Prahl et al., 1988; Bendle et al., 2005), and a thorough examination of the paleosalinlity proxy is best reserved for a small geographical boundary with a large salinity gradient and a shared temperatures regime. The Lake Qinghai region on the Tibetan plateau provides an ideal setting for this type of study, with Lake Qinghai and its surrounding

satellite lakes providing a salinity range between 1-76 ppt and a percent C37:4 range between 10-46% (Liu et al., 2008). Two studies have examined the salinity-%C37:4

111 relationship in the waters and surface sediments of this region, reporting a consistent

increase in C37:4 concentrations with decreasing salinities (Liu et al., 2008, 2011). These authors posit that a similar species of haptophyte may be responsible for alkenone production at these different sampling locations, and that salinity dictates the abundance

of C37:4 produced by the haptophyte. Based on their results, Liu et al. (2008) proposed a

C37:4-based paleosalinity proxy for application to lakes in the Qinghai region.

An examination of haptophyte community composition in the sediment paleorecord from Ace Lake, Antarctica revealed a clear shift in haptophyte species, and corresponding alkenone fingerprints, with changes in salinity (Coolen et al., 2004). With our study, we aim to verify if salinity or LCA-producing haptophyte species diversity is responsible for variations in alkenone signatures in a modern day lake environment. We examine a series of lakes on the Tibetan plateau that were previously the subject of alkenone-based paleosalinity studies (Chu et al., 2005; Liu et al., 2008) and use DNA sequencing to identify the haptophyte species present in these environments. Our study provides the first genetic confirmation of the haptophyte species responsible for alkenone

production in Lake Qinghai and the first genetic-based test of C37:4 concentrations across a salinity gradient in modern day lake environments.

2. Methods

2.1 Site description

The Lake Qinghai region on the Qinghai-Tibetan plateau in China is characterized by a semi-arid/arid and cold continental climate (Liu et al., 2011). The high altitude of

2180-3200m maintains a cool mean summer air temperature of 11-14°C (Shi et al., 1958;

112 Henderson et al., 2003; Colman et al., 2007). A negative precipitation-evaporation rate maintains the elevated salinity of lakes in the region. Lake Qinghai (Figure 1) is the largest inland brackish lake in China, with an average salinity of 12 ppt. Nearby lakes

Erhai and HaiyanWan have salinities of 0.97 ppt and 18, respectively (Table 1). Gahai lake is to the west, at the northeastern edge of the Qaidam Basin, with an elevated salinity of 76.4 ppt (Table 1).

Water sampling

We collected water samples in May of 2010. For each sample, we collected two liters of water using a Van Dorn water sampler and preserved one liter for DNA analyses and one liter for alkenone analyses. For alkenone analysis, we filtered one liter of water onto a pre-combusted (550°C) GF/F 0.7um, 47mm glass filter and kept it frozen at -20°C until freeze drying and extraction. For DNA analysis, we filtered one liter of water onto a

0.2um SterivexTM filter (Millipore, Billerica, MA, USA), flooded the filter with Puregene lysis buffer (Qiagen, Valencia, CA, USA) and froze it at -20°C until processing.

Lakes Erhai and HaiyanWan were sampled at the surface (Table 1). Lake Gahai was sampled at about 2m depth. Lake Qinghai was sampled twice at location QH2, at 3m

(QH2_3m) and at 6m depth (QH2_6m), and once at location QH12 at 8m depth

(QH12_8m) (Figure 2).

2.2 Alkenone extraction and analysis

We extracted the alkenone filters using three 20-minute bursts of sonication in dichloromethane (DCM), and ran the total lipid extracts on an Agilent 6890plus Gas

113 Chromatograph Flame Ionization Detector (GC-FID) for detection and quantification of alkenones using an internal C36 n-alkane standard and an external alkenone standard of

k known U37 -temperature value to ensure analytical precision (< 0.1°C). We used a Varian

VF200 60m GC column (60m × 250 µm × 0.10 µm) with the following temperature program: an initial temperature of 100°C (hold 1-min), ramp 20°C/min to 200°C (hold 1- min), ramp 4°C/min to 320°C (hold 0 min), hold at 320°C for 5-min.

2.3 DNA extraction and sequencing

We extracted SterivexTM filters using a Puregene Cell Kit (Qiagen) according to the manufacturer’s instructions. Genomic DNA was further purified using the Qiagen

DNA Purification to remove proteins and other contaminants that inhibit PCR reactions:

We quantified total extracted genomic DNA yields using a NanoDrop nucleic acid spectrophotometer (Thermo Scientific, Wilmington, DE).

We amplified genomic DNA using haptophyte-specific oligonucleotide (Coolen et al., 2004; Simon et al., 2000) primers targeting 18S rRNA coding regions. Forward and reverse primers corresponded to Escherichia coli 18S rRNA positions 429 and 887, respectively. Polymerase chain reactions (PCRs) were performed on an Eppendorf

Gradient Thermocycler (Hamburg, Germany) with the following conditions after

D'Andrea et al., (2006): 4 min initial denaturing at 96 °C, 35 cycles of denaturing for 30 s at 94 °C, followed by 40 s primer annealing at 55 °C and primer extension 40 s at 72 °C, with a final extension of 10 min at 72 °C. PCR reactions were run in 50 μl volume using

Promega GoTaq polymerase (Madison, WI, USA). Triplicate PCR products were pooled for each sample, and templates were purified using an Invitrogen PureLink Purification

114 kit (Carlsbad, CA, USA) with the high cut-off binding buffer to eliminate fragments <

200 bp. Purified PCR products were A-tailed and purified using the PureLink high cut-off kit. Cloning was performed using the Invitrogen TOP10 cloning kit with electro- competent cells according to the manufacturer's instructions. One hundred clones were picked for each sample. Plasmid DNA was isolated using a RevPrep Orbit robotic template preparation instrument (Genomic Solutions, Ann Arbor, MI), and prepared templates were sequenced on an ABI 3730XL (Applied Biosystems, Foster City, CA) capillary sequencer using the BigDye protocol with universal M13 forward and reverse primers according to the manufacturer's instructions. All sequencing was performed at the Marine Biological Laboratory W. M. Keck Ecological and Evolutionary Genetics

Facility. All sequences have been deposited in GenBank under accession numbers

XXXXX-XXXXXX.

2.4 Bioinformatics and phylogenetic reconstructions

A bioinformatics pipeline using the programs phred, cross-match, and phrap, translated chromatograms into base-calls and associated quality scores, removed vector sequences and assembled forward and reverse reads into full-length sequences for each of the cloned PCR amplicons (Ewing and Green, 1998; Lasek- Nesselquist et al., 2008).

Only sequences greater than 400 bp and with a complete forward and reverse primer were retained. Base-calls were verified and sequences were manually edited with the program

Consed (Gordon et al., 1998) for chromatogram viewing. We screened edited sequences for chimeras using the computer program Bellerophon (Huber et al., 2004) and the edited dataset contained 182 sequences. Assembled sequences were aligned using the ARB

115 software program v. 07.07.11 (Ludwig et al., 2004) against the July 2012 Silva 111 Ref database (Pruesse et al., 2007) using the FastAligner option followed by manual adjustment. Operational Taxonomic Units (OTUs) were obtained using UCLUST (Edgar,

2010) with a 97% cut-off criterion. OTU cutoffs of less than 97% failed to differentiate known reference taxa sequences. Consensus sequences for each OTU were selected for tree construction.

A custom filter of 429 positions was constructed manually for the Bayesian analyses of 18S rRNA sequences. We selected Cyclonexis annularis, Chrysoxys sp.,

Ochromonas danica, Odontella sinensis and Thraustochytrium multirudimentale as outgroups for our Bayesian analyses after de Vargas et al., (2007). We subjected our datasets to a Bayesian analysis using MrBayes version 3.0b4 (Ronquist and Huelsenbeck,

2003) under the GTR model of substitution (Lanave et al., 1984; Rodriguez et al., 1990;

Tavare,́ 1986) considering invariants and a gamma-shaped distribution of the rates of distribution among sites. The chain length for our analysis was 1,000,000 generations with trees sampled every 100 generations using MCMC (Markov Chain Monte Carlo) analysis. The first 10,000 trees were discarded as burn-in for the tree topology and posterior probability. OTU consensus sequences and full-length 18S rRNA gene sequences from reference taxa were analyzed to infer OTU species’ identities, and included a total of 185 taxa in our analysis. To compare microbial OTU distributions across samples, we used PRIMER-E statistical software (Clarke and Gorley, 2006) to perform cluster analyses and non-metric multidimensional scaling analyses (NMDS).

3. Results

116 3.1 Alkenone profiles

Hypersaline lake Gahai had the highest concentrations of alkenone lipids (Figure

4A), an order of magnitude higher than the other lakes. The concentrations of C37 alkenones for each of the six lakes samples are listed in Table 1. Some of our water samples had alkenone lipid concentrations that were below detection and we therefore include reference alkenone data for these samples from previous studies (Table 1). These supplemental alkenone data allow us to infer characteristic alkenone profiles although we interpret this data with caution.

Lake Erhai, the freshest lake study, had water concentrations dominated by C37:4 alkenones (Liu et al., 2011), with C37:4 comprising 92% of the C37 alkenones. Our lipid analysis did not detect alkenones in the Erhai water at the time of sampling. Lake

Qinghai samples 2_3m and 2_6m both had salinities of 12.62, and Qinghai sample

12_8m had a salinity of 12.61. Based on their GPS coordinates, both QH site 2 and site

12 were compared to corresponding water sampling locations from the summer sampling by Liu et al., (2011). Our results yielded alkenone signatures for Qinghai samples 2_3m

and 2_6m were both slightly dominated by the C37:4 alkenone, however the Liu et al.,

(2011) data for this location report a C37:3 dominant profile. Qinghai sample 12_8m did not yield detectable alkenones and those reported by Liu et al., had a dominant C37:4 profile. The Gahai lake sample was the most saline of the lakes sampled, with a salinity

of 76.4, and this lake had an alkenone profile dominated by C37:3, as expected for a saline lake.

3.2 Haptophyte DNA sequences

117 Our analyses yielded 182 haptophyte DNA sequences that passed quality screening, approximately 30 sequences per sample. These 182 haptophyte sequences were clustered at 97% similarity to yield 5 operational taxonomic units (OTUs). These 5

OTUs (China A-China E) were aligned to reference taxa to infer taxonomic identities

(Figure 3). Lake Erhai, the freshest lake, yielded haptophytes that clustered with

Chrysochromulina spp. and Prymnesium spp. Neither of these species is a known- alkenone producing haptophyte which corroborates our lipid analysis from this lake that failed to yield a discernable alkenone signature.

Lake Qinghai sample 2_3m yielded sequences from all five haptophyte OTUs.

The Qinghai sample from the same site but at 6m depth (2_6m) yielded a

Chrysochromulina spp. as well as a close relative of C. lamellosa (Figure 3). This result confirms the suspicion by Liu et al., (2008) that C. lamellosa was responsible for

alkenone production in Lake Qinghai due to the elevated C37:4 concentrations and lack of

C38 methyl ketones. The same pattern in OTU presence/absence in QH2_3m was observed in Qinghai sample 12_8m and Qinghai 2_6m and in Gahai (Table 2), the most saline sample. HaiyanWan, another brackish lake, only yielded sequences from

Chrysochromulina spp., although again we did not detect alkenones from this sample.

Our NMDS analysis of each lake’s haptophyte community composition revealed

the grouping of samples in accordance with C37:4 alkenone concentrations (Figure 4). The lakes with intermediate and high C37:4 concentration lakes, HaiyanWan, QH12_8m, and

Erhai, were most similar in OTU compositions (Figures 5). By comparing this with our

OTU distributions (Table 2), we can see that this similarity is shaped by the presence of

118 OTU D, the close relative of Chrysochrumlina spp. that was the dominant OTU in all three of these lake samples.

4. Discussion

4.1 Patterns in OTU distribution

Operational Taxonomic Unit representative sequence China A matches with

100% identity to OTU 5 from Theroux et al., (2010), an OTU that was identified in

Greenland lake BrayaSø sediments and associated with a haptophyte that makes

predominant C37:4 alkenone lipid signatures. OTU China A was discovered only in

Qinghai sample QH2_3m, and only a single DNA sequence that matched this OTU was

identified. Our alkenone analyses did reveal a dominant C37:4 alkenone signature for sample QH2_3m. However, OTU China A was not the only OTU detected in QH2_3m; in fact, this sample yielded sequences from all 5 OTUs. The only other putative alkenone producer from sample QH2_3m is a close relative to another OTU from a Greenland lake

(HundeSø) and a close relative of reference taxon C. lamellosa. Chrysotila lamellosa is a known lacustrine haptophyte species that has been isolated from Lake Xiarinur in China previously (Sun et al., 2007). Culture of the Lake Xiarinur C. lamellosa had a dominant

C37:3 signature, and lacked a methyl ketone (ibid). Our water sample had a dominant C37:4 alkenone signature and the presence of a methyl ketone, similar to the profile observed from lake BrayaSø (D’Andrea and Huang, 2005). Although both OTU China A and C. lamellosa-like OTU China B were detected in the QH2_3m sample, we believe OTU

China A was responsible for the C37:4-dominant alkenone fingerprint. A previous study in

Lake Qinghai (Liu et al., 2011) reported a C38:3 dominant alkenone profile near the location of Qinghai site 2 for the summer surface water sample (Table 1). This result

119 suggests that during our sampling of Qinghai site 2, the China A OTU was responsible for the majority of alkenone production but at other points during the year, the OTU

China B haplotype may be producing a C38:3 dominant alkenone signature.

Deeper in the water column at site QH2, at 6m depth, we only detected putative alkenone producing haptophyte, OTU China B, the close relative to C. lamellosa (Figure

3). This sample also had a dominant C37:4 alkenone signature, suggesting OTU China A concentrations were below detection at 6m depth. From sample QH2_8m we detected only a non-alkenone producing haptophyte related to Chrysochromulina spp. (OTU

China D), further indicating the change in haptophyte community with depth and the greater abundance of alkenone-producing haptophytes in the surface waters.

Lake Qinghai sample QH12_8m had sequences from OTU China B, as well as

OTU China D (Table 2). Of the three Lake Qinghai samples, QH2_3m had DNA sequences from all five China OTUs, whereas samples QH2_6m and QH12_8m had the presence of OTUs China B and China D. Samples QH2_3m and QH2_6m had dominant

C37:4 alkenone profiles, and while alkenones were below detection with our analysis, previous alkenone reports from site QH12 reported a dominant C37:4 alkenone signature.

As mentioned, previous results from site QH2 reported a C38:3 dominant alkenone signature, suggesting the potential seasonality in the dominance of alkenone producers at

this site, with alternations between in OTU A (dominant C37:4) and OTU B (dominant

C38:3).

The most saline site in this study, lake Gahai with a salinity of 76.4 ppt, yielded haptophyte sequences closely related to C. lamellosa-like OTU China B (Table 2), suggesting the plasticity of C. lamellosa in a variety of environmental salinities. Our

120 analysis revealed the dominance in the C38:3 alkenone, as expected because of the both the high salinity and the presence of C. lamellosa. Unsurprisingly, our two lakes that did not yield putative alkenone producing haptophyte sequences, Erhai and HaiyanWan, both did not yield detectable alkenone signatures. Both of these lakes have alkenones present in their surface and downcore sediments, suggesting the seasonal presence of alkenone- producing haptophytes in the photic zone of these lakes.

4.2 Implications for paleoclimatology

In this study, we examined haptophyte species assemblages as a means to explain gradients in alkenone lipid signatures across salinity gradients. As expected, our least

saline lake, Erhai, had the highest percentage of C37:4 alkenones, comprising greater than

90% of total C37 alkenones (Figure 4B). Our brackish sites, QH2 and QH12, both had

dominant C37:4 alkenones, although previously reported alkenone signature from QH2 at a similar salinity (Liu et al., 2011; Table 1) reported dominant C37:3 alkenone. HaiyanWan,

also a brackish lake, and Gahai, the most saline lake, had dominant C38:3 alkenone profiles with C37:4 percentages of about 43% and 39%, respectively (Figure 4B). These results were in agreement with previous studies from the Qinghai region demonstrating elevated

C37:4 concentrations in fresher waters (Liu et al., 2008; 2011).

Using the C37:4 percentages from our analyses, we calculated the approximate salinities as proposed by Liu et al., (2008), defined as %C37:4 = 53.4 (±7.8) – 1.73 (±0.45) x Salinity. Using this paleosalinity proxy, our approximated salinities based on %C37:4 were off by up to three orders of magnitude (Table 3). Given the conflicting alkenone lipid signatures reported in this study and by previous studies for the same lakes (and

121 salinities) in the Qinghai region (Liu et al., 2011), it is not surprising that the paleosalinity proxy as defined by Liu et al., (2008) failed to accurately approximate lake salinities. This highlights the significant intra-lake alkenone signature heterogeneity.

The lakes we sampled fall within the geographic bounds prescribed by Liu et al.,

(2008) and like previous studies in the region our samples demonstrated a relationship

between sample salinity and %C37:4 alkenones (Figure 7). While previous studies did not have water temperature measurements for comparison, our results confirm the C37:4- salinity relationship was not simply a result of a temperature effect, as there was no consistent trend between temperature and salinity (Figure S1). Liu et al. (2008) recommended only applying the paleosalinity proxy within a single lake system, an effort to avoid complications arising from multiple alkenone-producing haptophytes across disparate geographies. However, we now know that Lake Qinghai alone hosts five haptophyte species, and two distinct alkenone-producing haptophyte species, OTU A that

is likely responsible for elevated C37:4 concentrations, and OTU B responsible for C38:3

dominant alkenone signatures. Our results suggest that the observed variations in %C37:4 concentrations with salinity are in fact a result of a multiple-haptophyte system. It remains to be seen if other environmental factors besides salinity will preferentially select for different haptophyte populations.

5. Conclusions

The abundance of C37:4 alkenones is a hallmark of coastal and lacustrine environments. Concentrations of C37:4 alkenones can be affected by temperature, cell physiology, salinity, or as examined here, haptophyte taxonomy. Like previous studies in

122 the Qinghai region, our results showed a relationship between %C37:4 concentrations and salinity. Haptophyte C. lamellosa was isolated from nearby Lake Xiarinur (Sun et al.,

2007) and the lipid profiles from Lake Qinghai suggested the presence of C. lamellosa

(Liu et al., 2011). Our results confirmed the presence of C. lamellosa (OTU China B) in

Lake Qinghai, likely responsible for dominant C37:3 alkenone signatures. We also identified another known alkenone-producer, OTU China A, responsible for predominant

C37:4 alkenone signatures (D’Andrea et al., 2008; Theroux et al., 2010). Because of the individual alkenone-lipid signatures specific to these two haptophyte species, we believe changing environmental conditions preferentially select for different haptophyte species

resulting in the observed trends in C37:4 variations with salinity.

The dominant Isochrysidales (alkenone-producing) OTU across the lakes in this study, OTU B Chrysotila lamellosa, was present in the brackish samples from Qinghai and the most saline samples from Gahai. We did not detect putative alkenone-producing haptophytes from our samples in Erhai and HaiyanWan, likely reflecting the seasonality of alkenone production in the water column of these lakes. Future studies that capture the alkenone-producing haptophytes from Erhai, HaiyanWan, and the other lakes in the

Qinghai region will bolster the salinity-C37:4 relationship and help to resolve trends in C37:4 concentrations with haptophyte species. It will be important to verify that salinity alone, and not other environmental perturbations, are shaping the haptophyte community. Only then will the paleosalinity proxy prove valuable for application across a diversity of lake environments.

Acknowledgements

123 This work was supported by National Science Foundation awards to Y. Huang (EAR-

0902805, EAR-1122749) and L. Amaral-Zettler (EAR-1124192), a Brown SEED fund to

Y. Huang and L. Amaral-Zettler, and an American Association of University Women dissertation fellowship to S. Theroux.

124 References

Bendle J., Rosell-Mele ́ A. and Ziveri P. (2005) Variability of unusual distributions of alkenones in the surface waters of the Nordic seas. Paleoceanography 20, PA2001. doi:10.1029/ 2004PA001025.

Brassell SC, Eglinton G, Marlowe IT, Pflaumann U, Sarnthein M. (1986). Molecular stratigraphy - a new tool for climatic assessment. Nature 320:129-133.

Chu GQ, Sun Q, Li SQ, Zheng MP, Jia XX, Lu CF, Liu JQ, Liu TS. (2005). Long-chain alkenone distributions and temperature dependence in lacustrine surface sediments from China. Geochimica Et Cosmochimica Acta 69:4985-5003.

Clarke KR, Gorley RN. (2006). PRIMER v6: User Manual/Tutorial.

Colman SM, Yu S-Y, An ZH, Shen J, Henderson ACG. Late Cenozoic climate changes in China's western interior: a review of research on Lake Qinghai and comparison with other records. Quaternary Science Reviews 26:2281-2300.

Coolen MJL, Muyzer G, Rijpstra WIC, Schouten S, Volkman JK, Sinninghe Damsté J (2004). Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth and Planetary Science Letters 223:225-239.

Cranwell PA. (1985). Long-Chain Unsaturated-Ketones in Recent Lacustrine Sediments. Geochimica et Cosmochimica Acta 49:1545-1551.

D'Andrea WJ. (2008). Development and application of lacustrine alkenone paleothermometry in southwestern Greenland. Geological sciences PhD:197.

D'Andrea WJ, Huang Y. (2005). Long chain alkenones in Greenland lake sediments: Low δ13C values and exceptional abundance. Organic Geochemistry 36:1234-1241.

D'Andrea WJ, Lage M, Martiny JBH, Laatsch AD, Amaral-Zettler LA, Sogin ML, Huang YS. (2006). Alkenone producers inferred from well-preserved 18S rDNA in Greenland lake sediments. Journal of Geophysical Research-Biogeosciences 111:doi:10.1029/2005JG000121. de Vargas C, MP Aubry, I Probert, J Young (2007) Origin and evolution of coccolithophores: From coastal hunters to oceanic farmers In Evolution of Aquatic Photoautotrophs. P Falkowski, AH Knoll, eds. Elsevier Academic, New York, p. 251.

Edgar RC. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26:2460-2461.

125 Ewing B, Green P. (1998). Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Research 8:186-194.

Gordon D, Abajian C, Green P. (1998). Consed: A graphical tool for sequence finishing. Genome Research 8:195-202.

Henderson ACG, Holmes JA, Zhang JW, Leng MJ, Carvalho LR. A carbon- and oxygen- isotope record of recent environmental change from Qinghai Lake, NE TIbetan Plateau. Chinese Science Bulletin 48:1463-1468.

Huber T, Faulkner G, Hugenholtz P. (2004). Bellerophon: a program to detect chimeric sequences in multiple sequence alignments. Bioinformatics Bioinformatics 20:2317- 2319.

Lanave C, Preparata G, Saccone C, Serio G. (1984). A new method for calculating evolutionary substitution rates. Journal of Molecular Evolution 20:86-93.

Lasek-Nesselquist E, Bogomolni AL, Gast RJ, Welch DM, Ellis JC, Sogin ML, Moore MJ. (2008). Molecular characterization of Giardia intestinalis haplotypes in marine animals: variation and zoonotic potential. Diseases of Aquatic Organisms 81:39-51.

Li JG, Philp RP, Pu F, Allen J. (1996). Long-chain alkenones in Qinghai Lake sediments. Geochimica Et Cosmochimica Acta 60:235-241.

Liu WG, Liu ZH, Fu MY, An ZH. (2008). Distribution of the C37 tetra-unsaturated alkenone in Lake Qinghai, China: A potential lake salinity indicator. Geochimica Et Cosmochimica Acta 72:988-997.

Liu W, Liu Z, Wang H, He Y, Wang Z, Xu L. (2011). Salinity control on long-chain alkenone distributions in lake surface waters and sediments of the northern Qinghai- Tibetan Plateau, China. Geochimica et Cosmochimica Acta 75:1693-1703.

Liu Z, Henderson ACG, Huang Y. (2006). Alkenone-based reconstruction of late- Holocene surface temperature and salinity changes in Lake Qinghai, China. Geophys. Res. Lett. 33:L09707.

Ludwig W, Strunk O, Westram R, Richter L, Meier H, Yadhukumar Buchner A, Lai T, Steppi S, Jobb G, Forster W, Brettske I, Gerber S, Ginhart AW, Gross O, Grumann S, Hermann S, Jost R, Konig A, Liss T, Lussmann R, May M, Nonhoff B, Reichel B, Strehlow R, Stamatakis A, Stuckmann N, Vilbig A, Lenke M, Ludwig T, Bode A, Schleifer KH. (2004). ARB: a software environment for sequence data. Nucleic Acids Research 32:1363-1371.

Marlowe IT, Green JC, Neal AC, Brassell SC, Eglinton G, Course PA. (1984). Long- Chain (n-C37-C39) alkenones in the prymnesiophyceae. Distribution of alkenones and other lipids and their taxonomic significance. British Phycological Journal

126 19:203-216.

Mercer JL, Zhao MX, Colman SM. (2005). Seasonal variations of alkenones and U- 37(K) in the Chesapeake Bay water column. Estuarine Coastal and Shelf Science 63:675-682.

Müller PJ, Kirst G, Ruhland G, von Storch I, Rosell-Mele A. (1998). Calibration of the k' alkenone paleotemperature index U37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S). Geochimica Et Cosmochimica Acta 62:1757-1772.

Prahl FG, Muehlhausen LA, Zahnle DL. (1988). Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochimica Et Cosmochimica Acta 52:2303-2310.

Prahl FG, Wakeham SG. (1987). Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330:367-369.

Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig WG, Peplies J, Glockner FO. (2007). SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Research 35:7188-7196.

Rodriguez R, Oliver JL, Marin A, Medina JR. (1990). The general stochastic model of nucleotide substitution. Journal of Theoretical Biology 142:485-501.

Ronquist F, Huelsenbeck JP. (2003). MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics Bioinformatics 19:1572-1574.

Rosell-Melé A. (1998). Interhemispheric appraisal of the value of alkenone indices as temperature and salinity proxies in high-latitude locations. Paleoceanography 13:694-703.

Rosell-Melé A, Jansen E, Weinelt M. (2002). Appraisal of a molecular approach to infer variations in surface ocean freshwater inputs into the North Atlantic during the last glacial. Global and Planetary Change 34:143-152.

Schulz H-M, Schöner A, Emeis K-C. (2000). Long-chain alkenone patterns in the Baltic sea--an ocean-freshwater transition. Geochimica Et Cosmochimica Acta 64:469- 477. Seki O., Kawamura K., Sakamoto T., Ikehara M., Nakatsuka T. and Wakatsuchi M. (2005) Decreased surface salinity in the Sea of Okhotsk during the last glacial period estimated from alkenones. Geophys. Res. Lett. 32, L08710.

Shi YF, Lu MX, Li WZ. Physical geography with emphasis on geomorphology around Lake Qinghai. Acta Geogr. Sin. 24:33-48.

127

Sicre MA, Marty JC, Saliot A. (1990). n-Alkanes, fatty acid esters, and fatty acid salts in size fractionated aerosols collected over the Mediterranean Sea. J. Geophys. Res. 95:3649-3657.

Sicre M. A., Bard E., Ezat U. and Rostek F. (2002) Alkenone distributions in the North Atlantic and Nordic sea surface waters. Geochem. Geophys. Geosyst. 3. doi:10.1029/2001GC000159.

Simon N, Campbell L, Örnolfsdȯ ttiṙ E, Groben R, Guillou L, Lange M, Medlin LK. (2000). Oligonucleotide probes for the identification of three algal groups by dot blot and fluorescent whole-cell hybridization. The Journal of Eukaryotic Microbiology 47:76-84.

Sun Q, Chu G, Li S, Lü C, Zheng M. (2004). Long-chain alkenones in sulfate lakes and its paleoclimatic implications. Chinese Science Bulletin 49:2082-2086.

Sun Q, Chu GQ, Liu GX, Li S, Wang XH. (2007). Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellosa (Haptophyceae). Organic Geochemistry 38:1226-1234.

Tavaré S. (1986). Some probabilistic and statistical problems in the analysis of DNA sequences. Lectures on Mathematics in the Life Sciences 17:57-86.

Theroux SM, D'Andrea WJ, Toney JL, Amaral-Zettler LA, Huang Y. (2010). Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes: Implications for continental paleotemperature reconstructions. Earth and Planetary Science Letters 300:311-320.

Thiel V, Jenisch A, Landmann G, Reimer A, Michaelis W. (1997). Unusual distributions of long-chain alkenones and tetrahymanol from the highly alkaline Lake Van, Turkey. Geochimica Et Cosmochimica Acta 61:2053-2064.

Toney JL, Huang Y, Fritz SC, Baker PA, Grimm E, Nyren P. (2010). Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes in the interior United States. Geochimica Et Cosmochimica Acta

Volkman JK, Burton HR, Everitt DA, Allen DI. (1988). Pigment and lipid compositions of algal and bacterial communities in Ace Lake, Vestfold Hills, Antarctica. Hydrobiologia 165:41-57.

Wang RL, Zheng MP. (1998). Occurrence and environmental significance of long-chain alkenones in Tibetan Zabuye salt lake, SW China. International Journal of Salt Lake Research 6:281-302.

128 Table 1. Environmental characteristics of lakes in this study. ., ., ., ., et al et al et al et al et Liu Liu Liu Liu Reference study This study This study This study This study This study This 2011 GCA 2011 GCA 2011 GCA 2011 GCA 2011 nd nd nd 0.1 -0.9 -0.4 -0.4 -0.4 -0.4 0.100 Uk'37 nd nd nd 0.3 0.3 0.2 0.1 0.1 -0.2 -0.5 Uk37 :4 nd nd nd 0.5 0.5 0.4 92.5 29.7 54.4 42.9 %C37 nd nd nd 65.8 83.6 31.6 120.1 194.0 170.5 (ng/L) 2189.4 C37total nd nd nd 4.0 8.6 8.5 6.8 2.1 12.0 101.2 C37:2 nd nd nd 10.7 23.2 34.2 48.1 15.9 107.9 C37:3 1225.2 nd nd nd 33.9 40.8 50.7 65.3 13.5 863.1 179.4 C37:4 % DO 76.2 73.5 74.1 69.1 70.8 79.2 9.1 9.1 9.1 9.1 9.1 8.3 pH 1.89 21.11 21.12 21.06 30.39 104.7 Cond. mS/cm 13 8.8 9.8 (°C) 8.65 7.77 15.2 Temp Temp 19 24.5 24.5 Max Max site (m) site depth @ 0 8 1.9 (m) 3.02 5.95 0.15 sample Depth of 3190 3190 3195 3183 2848 Elevation 2011 ., 2011 99°52' 97°33' 36°32', 36°39', 36°39', 36°48', 36°55', 37°08', 100°43' 100°36' 100°36' 100°45' et al., et et al et Lat, Long 1.6 0.97 15.8 15.1 25.9 76.4 (ppt) 12.61 12.62 12.62 18.74 Salinity Salinity Erhai Erhai 2_3m 2_6m QH2* GaHai 12_8m 12_8m QH12+ Qinghai Qinghai Qinghai Qinghai

dete yeild not did when our analysis is provided studies previous from data Alkenone study. this in lakes for the samples of water 1. Descriptions Table HaiyanWan HaiyanWan Liu 8_2 from *Qinghai Liu 8_7 from +Qinghai 129

Table 2. OTU distribution among the six samples by percentage. Each of the 5

OTUs are listed with their closest relative reference taxon (based on BLAST distance) and the number of sequences that was detected in each sample for the specific OTU.

Qinghai Qinghai Qinghai Haiyan OTU OTU closest relative Erhai 2_3m 2_6m 12_8m Wan GaHai China A OTU 5/Greenland phylotype 2.9 China B Chrysotila lamellosa 2.9 28.1 3.2 82.8 China C Prymnesium spp. 10 40.0 China D Chrysochromulina spp. 90 20.0 71.9 96.8 100.0 17.2 China E Chrysochromulina spp. 34.3 Total seqence number 30 35 32 31 25 29

130 Table 3. Salinity reconstruction using Liu et al. (2008) paleosalinity proxy based on

%C37:4. Measured salinity for each lake is provided. Salinity reconstruction yielded values that were up to three orders of magnitude incorrect. Asterisk denotes when the data is from Liu et al., 2011.

Salinity %C37:4 Approximation Measured Salinity Erhai 179.40 3.39 1.60 Qinghai 2_3m 0.06 0.03 12.62 Qinghai 2_6m 0.08 0.03 12.62 Qinghai 2* 50.70 0.98 15.80 Qinghai 12* 65.30 1.26 15.10 HaiyanWan 13.50 0.29 25.90 GaHai 1.63 0.06 76.40

131

Figure 1. Map of lake locations in the Qinghai basin. Inset of Tibetan plateau with rectangle indicating the sampling locations (modified from Liu et al., 2011).

132

Figure 2. Water column profiles of sampling sites Qinghai 2 and Qinghai 12. Lakes were sampled on June 29, 2010. Arrows indicate location of samples QH2_3m,

QH2_6m and QH12_8m.

133

134 Figure 3. A consensus Bayesian phylogenetic tree depicting 18S rRNA gene-inferred relationships among haptophyte algae. An asterisk (*) indicates posterior probability values of 1.00; all other values as shown. Bold designates sequences from this study. Operational Taxonomic Units (OTUs) were defined at a 97% similarity level. Order classification after de Vargas et al. (2007), with number of sequences per order as indicated. The evolutionary distance for the number of changes per site is represented by the scale bar. GenBank accession numbers follow species and sequence names.

135 10000! C37:4!

! C37:3! C37:2! 1000!

100!

10! Alkenone concentraitons (ng/L) concentraitons Alkenone

1! Erhai! Qinghai Qinghai Qinghai 2*! Qinghai HaiyanWan! GaHai! 2_3m! 2_6m! 12*!

100! C37:4! 90! C37:3! C37:2! 80! 70!

! 60! 50!

Percent Percent 40! 30! 20! 10! 0! Erhai! Qinghai Qinghai Qinghai 2*! Qinghai HaiyanWan! GaHai! 2_3m! 2_6m! 12*!

Figure 4. Concentrations and percent abundances of alkenone homologs for each lake in this study. A. Alkenone concentrations plotted on a log scale. B. Relative alkenone concentrations. Asterisk denotes when the data is from Liu et al., 2011.

136

Figure 5. Non-metric multidimensional scaling plot of China lake samples and their

Bray-Curtis similarity in haptophyte OTU distributions. OTU distributions were

normalized by total and square root transformed. C37:4 percent concentrations are defined as high (> 40%), medium (20-50%) and low (< 20%). Boundaries of percent similarity as noted.

137

Figure 6. Hierarchical clustering dendrogram of haptophyte OTU distributions among lakes in this study based on Bray-Curtis similarity. OTU distributions were

normalized by total and square root transformed. C37:4 percent concentrations are defined as high (> 40%), medium (20-50%) and low (< 20%).

138

Figure 7. Relationship between lake water sample salinities and percent C37:4 alkenones in lakes from this study.

139

Figure S1. Relationship between lake salinity and water temperature for lakes in

this study. %C37:4 concentrations as shown.

140

CHAPTER 6

A PICOPLANKTON HAPTOPHYTE IS RESPONSIBLE FOR DOMINANT C37:4 ALKENONE PRODUCTION

SUSANNA THEROUX1,2

JAIME TONEY1,‡

ROBERT ANDERSEN3

YONGSONG HUANG1

LINDA AMARAL-ZETTLER1,2,4

1 BROWN UNIVERSITY, Department of Geological Sciences

2 MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

3 UNIVERSITY OF WASHINGTON, Friday Harbor Laboratories

4BROWN UNIVERSITY, Department of Ecology and Evolutionary Biology

‡ Present address: UNIVERSITY OF GLASGOW, School of Geographical and Earth Sciences

Prepared for:

Geochimica et Cosmochimica Acta

141

Abstract

Lacustrine alkenone records hold potential to be valuable sedimentary archives of continental paleotemperature. Uncertainties in alkenone biosynthesis by novel lake- dwelling haptophytes hamper the use of this proxy. All alkenone compounds are produced by haptophyte algae in the order Isochrysidales, a few species of which have been grown in culture to gauge alkenone production. The tetra-unsaturated alkenone is a signature compound found predominantly in lacustrine and coastal waters. Of all the cultured species, none has constitutively produced a dominant tetra-unsaturated alkenone

(C37:4) lipid signature. We recently maintained a dominant tetra-unsaturated alkenone producing haptophyte in enrichment culture. In an effort to isolate this organism, we size- fractionated the enrichment culture and grew these fractions at various temperatures to

k determine the alkenone-unsaturation index (U37 ) calibration. The < 3µm subculture was

the only alkenone-containing fraction, indicating this novel C37:4-predominant haptophyte

is a picoplankter. Our study is the first to successfully maintain the predominant C37:4-

k producing haptophyte in isolation and to determine its alkenone unsaturation index (U37 ) calibration. These results reveal the diverse ecology of alkenone-producing haptophytes

and provide an explanation for the widespread occurrence of C37:4-producing haptophytes in oligotrophic waters.

142 1. Introduction

k k' The alkenone unsaturation index (U37 and U37 ) is a valuable tool in marine paleothermometry. Marine alkenones are produced predominantly by ubiquitous haptophyte species Emiliania huxleyi and its close relative Gephyrocapsa oceanica,

k' allowing for a nearly global calibration of the U37 index based upon culture (Brassell et al., 1986; Prahl and Wakeham, 1987) and coretop calibrations (Muller et al., 1998). In the North Atlantic, and many coastal and lacustrine settings, the abundance of a

tetraunsaturated alkenone (C37:4) suggests the presence of a different alkenone-producing haptophyte species (Cranwell, 1985; Zink et al., 2001; Chu et al., 2005; D’Andrea and

Huang, 2005; Pearson et al., 2008; Theroux et al., 2010; Toney et al., 2010). Genetic studies have confirmed the presence of novel haptophyte species in lacustrine settings

(Coolen et al. 2004; D’Andrea et al., 2005; Theroux et al., 2010). Only coastal/lacustrine species Isochrysis galbana (Versteegh et al. 2001) and Chrysotila lamellosa (Rontani

2004; Sun et al., 2007) have been grown in culture to gauge their alkenone lipid

k production at different temperatures, a key step in empirically-deriving robust U37 calibrations. Isochrysis galbana and C. lamellosa both produce C37:4 alkenones although they are not the dominant alkenone isomer.

Lake George, ND hosts an annual haptophyte bloom event in late spring (Toney et al., 2010) that is responsible for the majority of annual alkenone production. The alkenones present in the sediments of Lake George can reach concentrations of 313 µg/g

TOC (Toney et al., 2010) and are characterized by the dominance of the C37:4 alkenone over the tri and di-unsaturated alkenones (C37:3 and C37:2, respectively). Using sediment enrichment culturing, Toney et al. (2012) were able to successfully maintain the

143 haptophyte responsible for C37:4 production (Hap-A) in enrichment culture, the first time this organism has ever been grown in a lab. The Hap-A species is a close relative of an

uncultured haptophyte from Ace Lake, Antarctica that produces a C37:4 dominant signature (Coolen et al., 2004, Theroux et al. 2010; Toney et al. 2012). Lake George also hosts a haptophyte species (Hap-B) closely related to C. lamellosa and Pseudoisochrysis

paradoxa that does not produce the dominant C37:4 alkenone (Theroux et al. 2010; Toney et al., 2012).

Only four haptophyte species have been successfully cultured to calibrate their

k U37 index, two marine species (E. huxleyi and G. oceanica) and two coastal/lacustrine species (C. lamellosa, I. galbana). Beyond these four species, the ecology of alkenone- producing haptophytes is unknown. Studies are mounting that suggest the novelty and unexplored diversity of lake-dwelling, alkenone-producing haptophytes (Coolen et al.,

2004; D’Andrea et al., 2006; Theroux et al., 2010). Recent work targeting marine haptophytes suggests there is an unknown haptophyte contribution to marine primary productivity, largely present in the picoplankton fraction (Liu et al., 2009). Formally defined as cells between 0.2-2µm in size, an extended definition of picoeukaryotes as any cell that passes through a 3µm filter is often applied in field studies (Moon-van der Staay et al., 2001) and is the definition we use here.

In an effort to further isolate and characterize lipid production by Hap-A, we performed an enrichment culturing experiment after Toney et al. (2012) and used this

source material to further isolate the C37:4-producing haptophyte. We performed a size fractionation experiment to isolate Hap-A and performed a series of manipulation experiments to gauge its alkenone response to temperature variation. This study is the

144 first to successfully visualize the novel species responsible for C37:4 production and to identify it as a picoplankter. Our contribution highlights the importance of molecular work to characterize haptophyte species that are previously unidentifiable given their small size and anonymous cell structure.

2. Methods

2.1 Enrichment culture and size fractionation subcultures

We initiated a sediment enrichment culture after Toney et al. (2012). We inoculated a glass beaker with 50 g Lake George surface sediment overlain with 2 liters of 0.2 μm filtered Lake George water amended with f/2 nutrients (Guillard, 1975). We placed this sediment enrichment culture in a 4°C incubator with a 24:0 light:dark light regime using a full-spectrum light. After four weeks, we subsampled the enrichment culture supernatant and separated this subculture into a < 3 μm, 3-5 μm, and a 5-10 μm fractions using Millipore Isopore cellulose nitrate filters (Billerica, Massachusetts). We filtered each size fraction onto a 44 mm GFF Whatman filter for alkenone analysis and kept the filter at -20°C until extraction. We detected alkenones in only the < 3μm micron fraction, and used this fraction for a temperature manipulation experiment.

We grew < 3 μm subcultures in glass vials in 0.2 μm filtered Lake George water amended with f/2 nutrients (Guillard, 1975). We grew the cultures in volumes of 50 ml at

5, 10, 15, 21 and 24°C under 24:0 light:dark regime using a full-spectrum light. All cultures were inoculated with approximately 8000 cells/ml. The cultures were sampled biweekly for cell counts and cell concentrations were determined using a haemocytometer to ensure the cultures were maintained in the logarithmic growth phase.

145 At the end of two months, we harvested the cultures for alkenone analysis by filtering onto a 44mm GFF Whatman filter and freezing the filter at -20°C until extraction. We also filtered a subsection of the < 3µm fraction onto a Sterivex (Millipore, Billerica, MA) filter and flooded the filter with Puregene cell lysis buffer (Qiagen, Hilden, Germany) to preserve for later DNA extraction and analysis. This Sterivex filter was also kept at -20°C until analysis.

2.2 Alkenone extraction and analysis

We extracted the alkenone filters using three 20-minute bursts of sonication in dichloromethane (DCM), and ran the total lipid extracts on an Agilent 6890plus Gas

Chromatograph Flame Ionization Detector (GC-FID) for detection and quantification of alkenones using an internal C36 n-alkane standard and an external alkenone standard of

k known U37 -temperature value to ensure analytical precision (< 0.1°C). We used a Varian

VF200 60m GC column (60 m × 250 μm × 0.10 μm) with the following temperature program: an initial temperature of 100°C (hold 1-min), ramp 20°C/min to 200°C (hold 1- min), ramp 4°C/min to 320°C (hold 0 min), hold at 320°C for 5-min. Analytical

k' accuracy, tracked via the injection of laboratory U37 standard of known temperature, was

k' ±0.042 U37 units (±0.1°C).

2.3 FISH experiment

We followed the protocol for Fluorescent in situ Hybridization (FISH) after

Toney et al. (2012). We concentrated five milliliters of the 5°C < 3μm culture by centrifugation and resuspended the cell pellet in -80% ethanol for three days at -20°C to

146 remove autofluorescence. We probed the cells using haptophyte-specific oligonucleotide probe PRYM02 (Simon et al., 2000) targeting the 18S rRNA gene sequence. We used specially designed oligonucleotide probes to target haptophyte “A”, the haptophyte present in the < 3 μm fraction. The sequence for this probe targeted basepairs 684 through 663 on the 18S rRNA molecule (5’-GCGCGTCCTTTTTCC-3’). We suspended cells hybridization buffer (18 µl 5M NaCl, 2ul 1M Tris–HCl pH 7.4, 1µl 1% SDS, 20µl

100% formamide, 1µl probe (0.2 nmol/µl), 58 µl distilled H2O) and the cells were incubated at 46°C for 2-h. We then centrifuged and resuspended the cells in wash buffer

(4.3µl 5M NaCl, 2 µl 1M Tris, 1µl 1% SDS, 92.7µl distilled H2O) and incubated at 48

°C for 15 min. We centrifuged the cells and resuspended them in wash buffer, mounted this onto an agar coated slide, and dried at room temperature for 1-h. Finally, we mounted the cells with Citiflor:Vectashield (4:1 v/v) (London) to preserve probe fluorescence.

Slides were viewed and photographed on a Zeiss Axioskop 2 MOT.

2.4 DNA extraction and sequencing

We extracted the DNA filters (Sterivex and Isopore) using a Puregene Cell Kit

(Qiagen) according to the manufacturer’s instructions. Genomic DNA was further purified using the Qiagen DNA Purification to remove proteins and other contaminants that inhibit PCR reactions. We quantified total extracted genomic DNA yields using a

NanoDrop nucleic acid spectrophotometer (Thermo Scientific, Wilmington, DE).

For capillary sequencing, we amplified genomic DNA using haptophyte-specific oligonucleotide (Coolen et al., 2004; Simon et al., 2000) primers targeting 18S rRNA coding regions. Forward and reverse primers corresponded to Escherichia coli 18S rRNA

147 positions 429 and 887, respectively. Polymerase chain reactions (PCRs) were performed on an Eppendorf Gradient Thermocycler (Hamburg, Germany) with the following conditions after D'Andrea et al. (2006): 4 min initial denaturing at 96 °C, 35 cycles of denaturing for 30 s at 94 °C, followed by 40 s primer annealing at 55 °C and primer extension 40 s at 72 °C, with a final extension of 10 min at 72 °C. PCR reactions were run in 50 μl volume using Promega GoTaq polymerase (Madison, WI, USA). Triplicate

PCR products were pooled for each sample, and templates were purified using an

Invitrogen PureLink Purification kit (Carlsbad, CA, USA) with the high cut-off binding buffer to eliminate fragments < 200 bp. Purified PCR products were A-tailed and purified using the PureLink high cut-off kit. Cloning was performed using the Invitrogen TOP10 cloning kit with electro- competent cells. Protocol was according to the manufacturer's instructions. One hundred clones were picked for each sample. Plasmid DNA was isolated using a RevPrep Orbit robotic template preparation instrument (Genomic

Solutions, Ann Arbor, MI), and prepared templates were sequenced on an ABI 3730XL

(Applied Biosystems, Foster City, CA) capillary sequencer using the BigDye protocol with universal M13 forward and reverse primers according to the manufacturer's instructions. All sequencing was performed at the Marine Biological Laboratory W. M.

Keck Ecological and Evolutionary Genetics Facility.

3. Results

3.1 Size fractionation experiment

The 5°C initial sediment enrichment culture was initially devoid of alkenones, but

after six weeks the enrichment culture supernatant yielded a C37:4 dominant alkenone

148 signature (Figure 1). This enrichment culture alkenone signature resembles those of

Lake George sediments and previous sediment enrichment cultures (Toney et al. 2010;

2012), with a dominant C37:4 isomer and the absence of a C38:3 methyl ketone. We separated the supernatant into three size fractions: < 3 μm, 3-5 μm, and 5-10 μm fractions. Of the three size fractions, only the < 3 μm size fraction yielded an alkenone signature, and this alkenone signature resembled the signature of the sediment enrichment culture (Figure 1) and Lake George surface sediments.

3.2 Alkenone production in < 3µm culture

Culture of the < 3 µm fraction yielded dominant C37:4 alkenones at 5°C, comprising about 68% of the total C37 alkenones (Table 1). This was on par with

concentrations of C37:4 that were observed in enrichment cultures (Toney et al., 2012) that ranged from 78% in the early stage of the experiment at 4°C to 47% at later stages during

the experiment (deep culture, Toney et al., 2010). The abundance of the C37:4 compound decreased as temperature increased, and 10, 15, 21 and 24°C cultures were all

characterized by a dominant C37:3 (Figure 2).

Alkenone concentrations per cell were similar to previous observations of alkenone concentrations in Isochrysis galbana (0.009-2pg/cell; Versteegh et al., 2001;

Marlowe et al., 1984), primarily about 0.2-2pg/cell. However, the 5°C culture had much higher alkenone concentrations, up to 8pg/cell. None of the cultures reached stationary phase by the termination of the experiment, and cultures between 10 and 24°C had double rates between 2-3 days (Table 1). The 5°C culture had the slowest growth rates with doubling times of greater than 7 days. The highest growth rates were observed at

149 15°C and lowest at 5°C. The lowest alkenone per cell concentrations were observed at

15°C and the highest at 5°C (Table 1).

The alkenones in the < 3µm fraction demonstrated a response temperature,

increasing in C37:3 alkenone concentrations with warmer temperatures (Figure 3). The

k degree of unsaturation of alkenone lipids at the different temperatures yielded a U37 calibration of

k U37 = 0.0373T – 0.7531

k where U37 is the concentration of (C37:2-C37:4)/(C37:2+C37:3+C37:4) and T is temperature in

k' Celsius. This calibration had a root mean squared error value of 1.88 °C. The U37

k calibration did not demonstrate a response to temperature (Figure 4). The U37

k calibration shared a similar slope to the U37 calibration for the Chrysotila lamellosa

k culture (Sun et al., 2007) and had a similar y-intercept to the German lakes U37

calibration (Zink et al., 2001). The C. lamellosa culture did not have a dominant C37:4

alkenone, although C37:4 alkenones were present. The Zink et al. (2001) calibration as well included data from lakes with dominant C37:4 alkenones.

3.3 Confirmation of Hap-A identity in < 3µm culture

To verify that alkenones were being produced in culture by Hap-A and only Hap-

A, we performed a FISH experiment using both a haptophyte-specific and a Hap-A specific probe. Using source material from the 5°C <3µm size fraction, the haptophyte specific probe PRYM02 only illuminated sub < 5µm sized cells (Figure 5A). These small cells were also illuminated with the Hap-A specific primer, confirming that the

150 only haptophytes in the < 3µm culture were Hap-A cells. To better observe the morphology of these cells, which is compromised when cells are preserved in ethanol as during the FISH protocol, we photographed the < 3µm fraction under DIC light (Figure

6). These cells are approximately 5µm and smaller, which can be expected if the cell grows after separation from the source material with the 3µm filter. The cells also appear to be engorged with lipid vacuoles.

4. Discussion

4.1 Cellular alkenone concentrations

The lowest growth rates and highest alkenone concentrations occurred at 5°C, in agreement with previous observations of E. huxleyi and G. oceanica cultures which had highest alkenone concentrations at lowest growth temperatures (Conte et al., 1998).

Alkenones are believed to be energy storage molecules in haptophytes (Epstein et al.,

2001; Eltgroth et al., 2005) and alkenone concentrations per cell increased during stationary growth phase and decrease after cultures are placed in the dark (Epstein et al.,

2001; Eltgroth et al., 2005). Our observed accumulation of alkenones at low temperatures and low growth rates may be the result of photosynthetic energy input exceeding the cell’s capacity for growth and division (Roessler, 1990) as was observed at both low temperatures and low nutrient conditions (Conte et al., 1998; Epstein et al., 2001).

The lowest growth temperature is also expected to have the highest C37:4

k concentration as this is the premise for the U37 alkenone unsaturation index. The accumulation of more unsaturated alkenones has previously been observed when haptophytes were grown under nutrient deplete conditions, and therefore lower growth rates (Prahl et al., 2003; Prahl et al., 2006). This is in agreement with what we observed,

151 with the highest C37:4 concentrations at the lowest growth rate at 5°C. The switch to C37:3 dominated alkenone signature at warmer temperatures may be a reflection of growth rate

and cell physiology, indicating that the C37:4 dominance is dependent on cell

k physiological status. The conservation of the U37 index over the five temperature experiments suggests that the culture is responding systematically to variations in temperature.

k 4.2 U37 index

k The U37 calibration for the < 3 µm culture had a y-intercept close to that of the

k' German lakes and a very similar slope to the C. lamellosa calibration. The U37 index failed to capture the alkenone response to temperature, indicating the importance of

incorporating the C37:4 alkenone when calibrating the Hap-A culture <3 µm culture.

Alkenone signatures were C37:4 dominant in the 5°C culture but C37:3 alkenones were dominant at warmer temperatures (Figure 2). Our FISH results confirm that Hap-A was the single haptophyte in the Hap-A cultures, and this culture served as the inoculum for warmer cultures. We did not perform a FISH experiment to confirm that Hap-A was

the only haptophyte in the warmer cultures, but the absence of a C38:3 methyl ketone in the warmer cultures is confirmation that the cultures were not contaminated with Hap-B cells

(Toney et al., 2012). Lake George enrichment cultures demonstrate a C37:4-dominant alkenone signature at 20°C (Toney et al., 2012) which our pure cultures failed to reproduce. The higher growth rates afforded by warmer temperatures may have resulted

152 in lower C37:4 alkenone production, as has previously been observed in fast growing E. huxleyi cultures (Prahl et al., 2003; Prahl et al., 2006).

4.3 Picohaptophytes

DNA sequencing revealed the dominant C37:4-producing haptophyte from Lake

George is a close relative of other dominant C37:4-producing, uncultured haptophytes from

Ace Lake, Antarctica (Theroux, et al., 2010; Coolen et al., 2004). Our discovery that the

Lake George Hap-A is of very small size suggests an ecological context for the occurrence of these unique alkenone producers. Picoplankton are often found in oligotrophic waters, their high surface area to volume ratio an advantage in low nutrient conditions (Vaulot et al. 2008). Hap-A was isolated from anoxic Lake George surface sediments (40m depth) and blooms in situ upon spring thaw and in culture when exposed to light (Toney et al., 2010; Toney et al., 2012). Hap-A’s high surface-area to volume ratio likely provides a competitive advantage to survive long periods of time out of the photic zone. Although some larger haptophytes can be mixotrophic (Brutemark and

Graneli, 2011), the small size of Hap-A cells likely prevents it from being mixotrophic

(Robert Andersen, personal communication).

Cyanobacteria compose a greater percentage of the picophototrophic microbial community, but the high cellular carbon concentrations and high growth rates of picoeukaryotes make them the largest fixers of carbon in the world’s oceans (Li, 1996;

Liu et al., 2009). Haptophytes account for up to 35% of the Atlantic Ocean picoeukaryote community (Not et al., 2005) and up to 25% of the global picophytoplankton community biomass (Cuvelier et al., 2010). Picoplankton haptophytes are known contributors to

153 primary production in ocean waters as indicated by microscopy, fluorescent in situ probing and pigment analyses (Not et al., 2005; Liu et al., 2009), but the majority of environmental haptophyte DNA sequences identify novel, uncultured species (Cuvelier et al., 2010). Picoplanktonic haptophytes can comprise 90% of the haptophyte community in marine waters (Cuvelier et al., 2010). Only six picoplanktonic haptophyte species are in culture, including species from genus Phaeocystis and Chrysochromulina, although none of these are alkenone-producers (Vaulot et al., 2008, Marlowe et al., 1984). Our paper provides the first description of an alkenone-producing pico-haptophytes.

Hap-A likely has a cyst hibernation stage that allows it to survive the long winters in Lake George (Toney et al., 2012). When haptophytes are incubated in the dark, they metabolize their alkenones beginning with the more-unsaturated compounds (Epstein et al., 2001). Similarly, when cells are grown in nutrient deplete conditions, they accumulate more unsaturated alkenones (Epstein et al., 2001; Eltgroth et al., 2005). The

abundance of C37:4 alkenones in Hap-A may serve as a competitive advantage, providing

large storage of readily metabolized compounds. High concentrations of C37:4 alkenones, in addition to a high surface area to volume ratio, may allow Hap-A cells to survive extended hibernation periods in anoxic waters, an advantage other haptophytes, including

Lake George Hap-B, do not have.

5. Conclusions

Our isolation and cultivation of the haptophyte Hap-A finally revealed that this

C37:4-dominant haptophyte is picoplanktonic, a characteristic that has likely allowed for its success in low-nutrient environments. The cultures of Hap-A displayed the

154 characteristic C37:4 dominant alkenone signature at 5°C but not at higher temperatures.

Fluorescence probing and the absence of a C38 methyl ketone confirmed that the Hap-A culture was in fact pure of other haptophyte contaminants and C37:3 dominance likely arises because of high growth rates. What we now know about the Lake George

k picohaptophyte has implications for the application of the U37 paleothermometry in lake

environments where dominant C37:4 alkenone signatures have been an enduring mystery.

The isolation and description of this novel species will further improve our understanding of haptophyte diversity and ecology in lake environments. Our results finally shed light on the ecology of a haptophyte that spends nine months of its year in the dark.

Acknowledgements

This work was supported by a National Science Foundation award to Y. Huang (EAR-

1122749) and L. Amaral-Zettler (EAR-1124192), a Brown SEED fund to Y. Huang and

L. Amaral-Zettler, and an American Association of University Women dissertation fellowship to S. Theroux.

155 References

Brassell SC, Eglinton G, Marlowe IT, Pflaumann U, Sarnthein M. (1986). Molecular stratigraphy - a new tool for climatic assessment. Nature 320:129-133.

Brutemark A, Graneli E. (2011). Role of mixotrophy and light for growth and survival of the toxic haptophyte Prymnesium parvum. Harmful Algae 10: 388-394.

Chu GQ, Sun Q, Li SQ, Zheng MP, Jia XX, Lu CF, Liu JQ, Liu TS. (2005). Long-chain alkenone distributions and temperature dependence in lacustrine surface sediments from China. Geochimica Et Cosmochimica Acta 69:4985-5003.

Conte MH, Thompson A, Lesley D, Harris RP. (1998). Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica. Geochimica Et Cosmochimica Acta 62:51-68.

Coolen MJL, Muyzer G, Rijpstra WIC, Schouten S, Volkman JK, Sinninghe Damsté J. (2004). Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth and Planetary Science Letters 223:225-239.

Cranwell PA. (1985). Long-Chain Unsaturated-Ketones in Recent Lacustrine Sediments. Geochimica Et Cosmochimica Acta 49:1545-1551.

Cuvelier ML, Allen AE, Monier A, McCrow JP, Messie M, Tringe SG, Woyke T, Welsh RM, Ishoey T, Lee J-H, Binder BJ, DuPont CL, Latasa M, Guigand C, Buck KR, Hilton J, Thiagarajan M, Caler E, Read B, Lasken RS, Chavez FP, Worden AZ. (2010). Targeted metagenomics and ecology of globally important uncultured eukaryotic phytoplankton. Proceedings of the National Academy of Sciences 107:14679-14684.

Cuvelier ML, Allen AE, Monier A, McCrow JP, Messie M, Tringe SG, Woyke T, Welsh RM, Ishoey T, Lee J-H, Binder BJ, DuPont CL, Latasa M, Guigand C, Buck KR, Hilton J, Thiagarajan M, Caler E, Read B, Lasken RS, Chavez FP, Worden AZ. (2010). Targeted metagenomics and ecology of globally important uncultured eukaryotic phytoplankton. Proceedings of the National Academy of Sciences 107:14679-14684.

D'Andrea WJ, Huang Y. (2005). Long chain alkenones in Greenland lake sediments: Low δ13C values and exceptional abundance. Organic Geochemistry 36:1234-1241.

D'Andrea WJ, Lage M, Martiny JBH, Laatsch AD, Amaral-Zettler LA, Sogin ML, Huang YS. (2006). Alkenone producers inferred from well-preserved 18S rDNA in Greenland lake sediments. Journal of Geophysical Research-Biogeosciences 111:doi:10.1029/2005JG000121.

156

Eltgroth ML, Watwood RL, Wolfe GV. (2005). Production and cellular localization of neutral long-chain lipids in the haptophyte algae Isochrysis galbana and Emiliania huxleyi. Journal of Phycology 41:1000-1009.

Epstein BL, D'Hondt S, Hargraves PE. (2001). The possible metabolic role of C-37 alkenones in Emiliania huxleyi. Organic Geochemistry 32:867-875.

Guillard RRL (1975). Culture of phytoplankton for feeding marine invertebrates. In Culture of marine invertebrate animals. WL Smith, MH Chanley, eds. Plenum Press, New York, p. 26.

Li WKW. (1995). Composition of ultraphytoplankton in the central North Atlantic. Mar Ecol Prog Ser 122:1–8.

Liu H, Probert I, Uitz J, Claustre H, Aris-Brosou S, Frada M, Not F, de Vargas C. (2009). Extreme diversity in noncalcifying haptophytes explains a major pigment paradox in open oceans. Proceedings of the National Academy of Sciences of the United States of America 106:12803-12808.

Moon-van der Staay SY, van der Staay GWM, Guillou L, Vaulot D, Claustre H & Medlin LK (2000) Abundance and diversity of prymnesiophytes in the picoplankton community from the equatorial Pacific Ocean inferred from 18S rDNA sequences. Limnol Oceanogr 45: 98–109.

Marlowe IT, Green JC, Neal AC, Brassell SC, Eglinton G, Course PA. (1984). Long- Chain (n-C37-C39) alkenones in the prymnesiophyceae. Distribution of alkenones and other lipids and their taxonomic significance. British Phycological Journal 19:203-216.

Muller̈ PJ, Kirst G, Ruhland G, von Storch I, Rosell-Mele A. (1998). Calibration of the k' alkenone paleotemperature index U37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S). Geochimica Et Cosmochimica Acta 62:1757-1772.

Not F, Massana R, Latasa M, Marie D, Colson C, Eikrem W, Pedros-Alios C, Vaulot D, Simon N. (2005). Late summer community composition and abundance of photosynthetic picoeukaryotes in Norwegian and Barents Seas. Limnology and Oceanography 50:1677-1686.

Pearson EJ, Juggins S, Farrimond P. (2008). Distribution and significance of long-chain alkenones as salinity and temperature indicators in Spanish saline lake sediments. Geochimica Et Cosmochimica Acta 72:4035-4046.

Prahl FG, Wolfe GV, Sparrow MA. (2003). Physiological impacts on alkenone paleothermometry. Paleoceanography 18:doi:10.1029/2002PA000803.

157

Prahl FG, Muehlhausen LA, Zahnle DL. (1988). Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochimica Et Cosmochimica Acta 52:2303-2310.

Prahl FG, Wakeham SG. (1987). Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330:367-369.

Prahl FG, Mix AC, Sparrow MA. (2006). Alkenone paleothermometry: Biological lessons from marine sediment records off western South America. Geochimica et Cosmochimica Acta 70:101-117.

Riebesell U, Zondervan I, Rost B, Tortell PD, Zeebe RE, Morel FMM. (2000). Reduced calcification of marine plankton in response to increased atmospheric. Nature 407:364-367.

Roessler PG. (1990). Environmental control of glycerolipid metabolism in microalgae: commercial implications and future research directions. Journal of Phycology 26:393–399.

Rontani JF, Beker B, Volkman JK. (2004). Long-chain alkenones and related compounds in the benthic haptophyte Chrysotila lamellosa Anand HAP 17. Phytochemistry 65:117-126.

k' Sikes EL, Volkman JK. (1993). Calibration of alkenone unsaturation ratios (U37 ) for paleotemperature estimation in cold polar waters. Geochimica et Cosmochimica Acta 57:1883-1889.

Sun Q, Chu GQ, Liu GX, Li S, Wang XH. (2007). Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellosa (Haptophyceae). Organic Geochemistry 38:1226-1234.

Theroux SM, D'Andrea WJ, Toney JL, Amaral-Zettler LA, Huang Y. (2010). Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes: Implications for continental paleotemperature reconstructions. Earth and Planetary Science Letters 300:311-320.

Toney JL, Huang Y, Fritz SC, Baker PA, Grimm E, Nyren P. (2010). Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes in the interior United States. Geochimica Et Cosmochimica Acta

Toney JL, Theroux S, Andersen RA, Coleman A, Amaral-Zettle L, Huang Y. (2012).

Culturing of the first 37:4 predominant lacustrine haptophyte: Geochemical, biochemical, and genetic implications. Geochimica et Cosmochimica Acta 78:51 - 64.

158 Vaulot D, Eikrem W, Viprey M, Moreau H. (2008). The diversity of small eukaryotic phytoplankton (≤3 μm) in marine ecosystems. FEMS Microbiology Reviews 32:795-820.

k' Versteegh GJM, Riegman R, de Leeuw JW, Jansen JHF. (2001). U37 values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations. Organic Geochemistry 32:785-794.

Zink KG, Leythaeuser D, Melkonian M, Schwark L. (2001). Temperature dependency of long-chain alkenone distributions in Recent to fossil limnic sediments and in lake waters. Geochimica Et Cosmochimica Acta 65:253-265.

159 Table 1. Alkenone abundance data for < 3µm culture experiment.

Growth Final cell rate concentration (Doubling C37:4 C37:3 C37:2 C37total % C37alk/ Temp. (°C) (cell/ml) time) (ng/L) (ng/L) (ng/L) (ng/L) C37:4 cell (pg) Uk37 Uk'37 5 60000 7.22 6929 2636 491 10056 68.9 0.17 -0.64 0.16 10 800000 3.16 7876 15465 0 23341 33.7 0.03 -0.34 0.00 15 5120000 2.25 5139 16337 2813 24288 21.2 0.00 -0.10 0.15 21 640000 3.32 2119 27009 1586 30713 6.9 0.05 -0.02 0.06 24 800000 3.161 958 18385 3612 22956 4.2 0.03 0.12 0.16

160 Figure 1. Gas chromatograms of culture size fractions. Enrichment culture was subdivided into < 3µm , 3-5µm, and a 5-10µm fractions. Enrichment culture chromatogram is at 6 weeks.

161 100!

80! ! 60! 37:4 Area!

Percent 40! 37:3 Area! 20! 37:2 Area!

0! 5! 10! 15! 21! 24! Temperature (°C)!

Figure 2. Percent abundances of < 3µm culture alkenones at all temperatures.

162

C37:3! A ! B < 3µm 5°C! C37:4! < 3µm 21°C! ! ) pA

C37:3! C38:3et! !

C38:3et!

Response ( C C 38:2! C38:4! 38:2! C37:2! C37:4! C37:2! C38:4!

Retention time (min)! ! Retention time (min)!

Figure 3. Abundance of C37 isomers in the < 3µm cultures. A. GC-FID chomatogram of < 3µm culture at 5°C culture. B. GC-FID chromatogram of <3µm culture at 21°C.

163

A 0.3! Uk'37 = 0.001T + 0.09! . R" = 0.01121! 0.2! 0.1! ! 0! -0.1! -0.2! -0.3! Uk37 = 0.0373T - 0.7531! Uk37 and Uk'37 -0.4! R" = 0.9443! -0.5! -0.6! -0.7! 0! 5! 10! 15! 20! 25! 30! Temperature (°C)!

B. 1.2! 1! 0.8! <3um culture! 0.6! C. lamellosa!

! 0.4! PolarMarine! 0.2! P. paradoxa! Uk37 0! GermanLakes! -0.2! Marine! -0.4! Lake George enrichment! -0.6! Linear (<3um culture)! -0.8! 0! 5! 10! 15! 20! 25! 30! Temperature (°C)!

k Figure 4. Alkenone unsaturation indices for the < 3µm fraction culture A. U37 and

k' k k U37 calibrations. The RMSE for the < 3µm U37 calibration is 1.88°C. B. The U37 index for both the <3um culture and previously published calibrations. References for the calibrations are: C. lamellosa (Sun et al., 2007), Marine Polar (Sikes and

Volkman 1993), P. paradoxa (Theroux, Chapter 4), German lakes (Zink et al. 2001),

Marine calibration (Prahl et al., 1998), Lake George enrichment (Toney et al., 2012).

164

Figure 5. Lake George enrichment culture < 3µm fraction hybridized with haptophyte and Lake George Hap-A specific probes. A1) Light micrograph of enrichment culture hybridized with haptophyte-specific oligonucleotide probe. A2)

Same as A1 under blue light. Arrows denote fluorescing cells. B1) Light micrograph of enrichment culture with Lake George Hap-A specific probe. B2) Same as B1 with blue light. Arrows denote fluorescing cells.

165

Figure 6. Photomicrograph of < 3µm culture. Cells appear to be engorged with lipid vacuoles.

166 CHAPTER 7

LATE BLOOMERS: IN SITU and EX SITU OBSERVATIONS OF HAPTOPHYTE

SPECIES DISTRIBUTIONS THROUGHOUT THE COURSE OF A SEASONAL

BLOOM

SUSANNA THEROUX1,2

JAIME L. TONEY3

ROBERT ANDERSEN4

PAUL NYREN5

RICK BOHN5

JEFFREY SALACUP1

YONGSONG HUANG1

LINDA AMARAL-ZETTLER1,2,6

1BROWN UNIVERSITY, Department of Geological Sciences 2MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center 3UNIVERSITY OF GLASGOW, School of Geographical and Earth Sciences 4UNIVERSITY OF WASHINGTON, Friday Harbor Laboratories 5NORTH DAKOTA STATE UNIVERSITY, Central Grasslands Research Extension Center 6BROWN UNIVERSITY, Department of Ecology and Evolutionary Biology

Prepared for: Geochimica et Cosmochimica Acta

167

Abstract

k The ability to reconstruct paleotemperatures using the alkenone-based U37

k temperature proxy is dependent upon the selection of a proper U37 calibration. When multiple haptophyte species occur in a single environment, the selection of an appropriate

k species-specific U37 calibration is unclear. Previous research revealed the presence of two haptophyte species in Lake George, ND and the contribution to the alkenone sediment record by each phylotype was unknown. We analyzed the haptophyte community during the spring bloom event in Lake George, ND to determine the haptophyte (or haptophytes) responsible for the seasonal alkenone production. We report the first study to monitor a lacustrine haptophyte bloom throughout the course of its seasonal cycle using a combined approach of DNA sequencing, quantitative PCR and geochemical analyses. Our results revealed a seasonal succession in haptophyte species, with an early and late alkenone peak dominated by different haptophyte species. Culture- based and in situ analyses revealed individual alkenone fingerprints for each haptophyte species, although at times the water column alkenones were composite signatures from

k both species. Our results established a suite of U37 calibrations that can be used in consortium to accurately reconstruct paleotemperatures from Lake George sediments.

This multidisciplinary approach can serve as a guideline for future alkenone-based temperature reconstructions from multiple haptophyte environments.

168 1. Introduction

Since its establishment in 1986 (Brassell et al., 1986), the alkenone-based

k k' paleotemperature proxy (U37 and U37 ) has been a popular and reliable tool for reconstructing sea surface temperatures from marine sediment records. The desire to extend the use of this paleotemperature proxy to lake sediments resulted in an increased

k interest in lacustrine alkenone production. The U37 proxy relies upon the degree of unsaturation of alkenone compounds produced exclusively by haptophyte algae in the order Isochrysidales. Marine alkenones are predominantly produced by cosmopolitan haptophyte species Emiliania huxleyi and its close relative Gephyrocapsa oceanica. Their

k' predominance in marine waters allows for the application of a single, universal U37 calibration, one that was first developed empirically though culture work (Brassell, 1986;

Prahl and Wakeham 1987) and later confirmed with a global core-top survey (Müller et al.,1998).

Like their marine counterparts, lacustrine alkenone unsaturation ratios correspond to lake seasonal surface water temperatures and mean annual air temperature (Chu et al.,

2005; D'Andrea and Huang, 2005; Sun et al., 2007; Zink et al., 2001). However, the novelty and unexpected diversity of lacustrine haptophyte species has impeded the application of this proxy to lacustrine sediments. Only two species of coastal/lacustrine alkenone-producing haptophytes are available in culture, Chrysotila lamellosa and

Isochrysis galbana (Versteegh et al. 2001; Rontani 2004; Sun et al., 2007). In lakes such as Lake George, ND, the presence of multiple haptophyte species complicates the

k selection of a U37 calibration (Theroux et al., 2010; Toney et al., 2012).

169 Traditional patterns in alkenone distribution differ between marine and lacustrine

environments. Marine sediments typically have low concentrations of the C37:4 alkenones and the presence of the C38:3 methyl ketone while in contrast, lacustrine alkenone profiles are characterized by abundant tetraunsaturated alkenone isomers (C37:4) and the absence of a C38:3 methyl ketone (Chu et al., 2005; Liu et al., 2008; Pearson et al., 2008).

Haptophyte species responsible for alkenone production in lakes were once inferred from alkenone lipid signatures, but DNA sequencing has revealed a more complex species history in many locations (Coolen et al., 2004; D’Andrea and Huang, 2005; Theroux et al., 2010). The discovery of a novel clade of haptophytes from a series of lakes in

Greenland (D’Andrea et al., 2006; Theroux et al., 2010) highlights the potentially untapped diversity of haptophyte species in lakes. Lacustrine haptophytes experience bloom events (D’Andrea et al., 2011; Toney et al., 2010) that result in large inputs of alkenones to the lake sediment record. The environmental triggers of seasonal alkenone- producing haptophyte blooms are unknown, although nutrient loading and seasonal irradiance levels can cause marine bloom events (Tyrrell and Merico, 2004).

Alkenone production in Lake George, North Dakota is attributed to two haptophyte species whose DNA was recovered from the lake’s surface sediments. Hap-A is a close relative of haptophytes from Ace Lake, Antarctica (Coolen et al., 2004) and cultured species Isochrysis galbana (Theroux et al., 2010). Hap-B is a close relative of cultured haptophytes Pseudoisochrysis paradoxa and Chrysotila lamellosa (Theroux et al., 2010). Lake George is situated in the Northern Great Plains and has a pristine alkenone sediment record that reflects the lake’s environmental change over the past

8000 years (Toney, 2011). Climate reconstructions from tree rings and speleotherms in

170 this region are discontinuous or scarce (Toney, 2011), emphasizing the importance of a reliable temperature proxy.

The accurate estimation of paleotemperature from the Lake George sediment record requires an improved understanding of both the ecology of these novel haptophyte species and their modes of alkenone production. In particular, our study aimed to determine a) if both haptophyte species are responsible for alkenone production and b) if both species produce similar alkenone signatures. To address these questions, we sampled the 2011 Lake George haptophyte bloom event and analyzed the geochemical and biological artifacts of the bloom. We used high-throughput DNA sequencing, quantitative PCR and culture manipulations to gauge haptophyte succession and alkenone production, three approaches that yielded corroborating results. Our study provides the first comprehensive depiction of species succession during a haptophyte bloom and uncovers how the haptophyte seasonal cycle is recorded in the lake sedimentary record.

2. Methods

2.1 Site description

Lake George (46.74°N, 99.49°W) is situated in the Northern Great Plains in North

Dakota (Figure 1). Lake George has a maximum depth of 60m and a salinity averaging

9.72 ppm, making it unique among the lakes in North Dakota (Fritz, 2011). Glacial retreat formed the lake over 12,000 years ago, and the lake’s salinity is likely maintained by input from deep sources of saline groundwater (ibid). Lake George sits atop the Hell

Creek, Fox Hills and Pierre formations of the Upper Cretaceous (Whitehead, 1996), and

171 sources its high sulfate concentrations (7370 mg/L) from the Pierre shale. The average pH of the lake waters is 8.98 (Toney et al., 2010).

2.2 Bloom sampling

In April 2011 we began sampling the waters of Lake George to gauge haptophyte cell and alkenone abundance throughout the course of the seasonal cycle. Biweekly sampling of Lake George continued until late July. We collected water using a VanDorn water sampler at 0, 5, and 10m depths. From each depth, we collected 1L each for DNA analysis, alkenone analysis, for geochemical analysis. For alkenone analysis, we filtered one liter of water onto a pre-combusted (550°C) GF/F 0.7µm, 47 mm glass filter, and kept it frozen until analysis. For DNA analysis, we filtered a separate liter of lake water onto a 0.2µm SterivexTM filter (Millipore, Billerica, MA, USA), flooded the filter with

Puregene lysis buffer (Qiagen, Valencia, CA, USA), and froze it at -20°C until processing. Water collected for geochemical analysis was filtered through a 0.2 μm filter to remove microbes and particulate matter and stored at -20°C until processing.

2.3 Geochemical analysis

We used a YSI 85 meter to measure lake water temperature, pH, dissolved oxygen and conductivity in situ. We measure nitrate, nitrite and phosphate using a

Westco Smartchem 200 Discrete Analyzer at Brown University. Trace metals were analyzed at the Barnstable County Department of Health and Environment Laboratory using Agilent ICP-MS EPA method 200.8 and a Perkin Elmer Flame Atomic

Absorption Spectrometer (Ca, Mg, Na).

172 2.4 Mock-bloom event

We initiated a sediment enrichment culture after Toney et al. (2011). We inoculated a glass beaker with 50g Lake George surface sediment overlain with 2 liters of

0.2 µm filtered Lake George water amended with f/2 nutrients (Guillard 1975). We placed this sediment enrichment culture in a 4°C incubator with a 24:0 light:dark light regime using a full-spectrum light. We performed biweekly sampling of the enrichment culture supernatant to gauge cell numbers and collect water for DNA and alkenone analysis. We filtered 50 ml onto a 44 mm GFF Whatman filter for alkenone analysis and kept the filter at -20°C until extraction. We filtered 15 ml onto a 25 mm 0.2 µm Isopore

(Millipore) cellulose nitrate filter for DNA analysis and flooded the filter with lysis buffer and stored at -20°C until extraction.

k 2.5 Culture studies to determine Hap-B U37 calibration

We achieved individual clonal cultures of Haptophyte B (#A17-903) through cell- picking with sterilized glass pipettes. We grew these clonal cultures in glass vials in

0.2μm filtered Lake George water amended with f/2 nutrients (Guillard 1975). We grew the cultures at 5,10,15,21 and 24°C under 24:0 light:dark regime using a full-spectrum light. The cultures were sampled biweekly for cell counts and cell concentrations were determined using a haemocytometer to ensure the cultures were maintained in the logarithmic growth phase. At the end of two months, we harvested the cultures for alkenone analysis.

2.6 Alkenone extraction and analysis

173 We extracted the alkenone filters using three 20-minute bursts of sonication in dichloromethane (DCM), and ran the total lipid extracts on an Agilent 6890plus Gas

Chromatograph Flame Ionization Detector (GC-FID) for detection and quantification of alkenones using an internal C36 n-alkane standard and an external alkenone standard of

k known U37 -temperature value to ensure analytical precision (< 0.1°C). We used a Varian

VF200 60m GC column (60m × 250μm × 0.10 μm) with the following temperature program: an initial temperature of 100°C (hold 1-min), ramp 20°C/min to 200°C (hold 1- min), ramp 4°C/min to 320°C (hold 0 min), hold at 320°C for 5-min.

Alkenones were extracted from the sediment and/or filters and isolated from co- eluting compounds before calculation of alkenone concentrations and ratios. Sediments and filters were freeze-dried before analysis. Samples were then powdered and extracted three times with dichloromethane on an Accelerated Solvent Extractor (Dionex, ASE200) at 150°C and 1500 psi to produce a total lipid extract (TLE). TLEs were then separated over silica gel using hexane, dichloromethane, and methanol to yield hydrocarbon, ketone

(alkenone), and polar fractions, respectively.

Two quantification standards (n-C36 and n-C37 alkanes) were added to all samples before being injected from an autosampler into a 112°C CIS-PTV (cooled injection system-programmed temperature vaporizer) inlet operated in solvent vent mode.

After the initial vent, the inlet was ramped at 12°C/min to 240°C, held isothermally for 5 minutes, ramped again at 12°C to 320°C, and held isothermally for 2 minutes before cryogenic cooling. A 60m, 0.32mm ID, 0.10um filmDB-1 with a 5m fused guard column

(DB-1 duraguard) was used. The oven began at 90°C for 2 minutes, was ramped at

40°C/min to 255, at 1°C to 302, and at 10°C to 325 where it was held isothermally for 20

174 minutes. Hydrogen was used as a carrier gas. Analytical accuracy, tracked via the

k' k' injection of laboratory U37 standard of known temperature, was ±0.042 U37 units

(±0.1°C).

2.7 DNA extraction and sequencing

We extracted the DNA filters (Sterivex and Isopore) using a Puregene Cell Kit

(Qiagen) according to the manufacturer’s instructions. Genomic DNA was further purified using the Qiagen DNA Purification to remove proteins and other contaminants that inhibit PCR reactions: We quantified total extracted genomic DNA yields using a

NanoDrop nucleic acid spectrophotometer (Thermo Scientific, Wilmington, DE).

For capillary sequencing, we amplified genomic DNA using haptophyte-specific oligonucleotide (Coolen et al., 2004; Simon et al., 2000) primers targeting 18S rRNA coding regions. Forward and reverse primers corresponded to Escherichia coli 18S rRNA positions 429 and 887, respectively. Polymerase chain reactions (PCRs) were performed on an Eppendorf Gradient Thermocycler (Hamburg, Germany) with the following conditions after D'Andrea et al. (2006): 4min initial denaturing at 96 °C, 35 cycles of denaturing for 30 s at 94 °C, followed by 40 s primer annealing at 55 °C and primer extension 40 s at 72 °C, with a final extension of 10 min at 72 °C. PCR reactions were run in 50μl volume using Promega GoTaq polymerase (Madison, WI, USA). Triplicate

PCR products were pooled for each sample, and templates were purified using an

Invitrogen PureLink Purification kit (Carlsbad, CA, USA) with the high cut-off binding buffer to eliminate fragments < 200 bp. Purified PCR products were A-tailed and purified using the PureLink high cut-off kit. Cloning was performed using the Invitrogen TOP10

175 cloning kit with electro- competent cells. Protocol was according to the manufacturer's instructions. One hundred clones were picked for each sample. Plasmid DNA was isolated using a RevPrep Orbit robotic template preparation instrument (Genomic

Solutions, Ann Arbor, MI), and prepared templates were sequenced on an ABI 3730XL

(Applied Biosystems, Foster City, CA) capillary sequencer using the BigDye protocol with universal M13 forward and reverse primers according to the manufacturer's instructions. All sequencing was performed at the Marine Biological Laboratory W. M.

Keck Ecological and Evolutionary Genetics Facility.

For Ion Torrent (IT) sequencing, we performed genomic DNA amplifications using eukaryotic-specific primers targeting the 18S rRNA gene variable V9 region

(1380F, 1389F, 1510R) Amaral-Zettler et al., 2009). We then ligated both barcodes and adapters to these V9 amplicons. We sequenced the amplicons using an Ion Torrent

Personal Genome Machine (PGM) with bidirectional sequencing at the Life Technologies facility.

2.8 Bioinformatics

Ion Torrent DNA sequences reads were trimmed and screened for quality after

Huse et al. (2007), including the paired end information as a measure of accuracy. We clustered the sequences at 97% similarity using SLP-PWAN (Huse 2010) into

Operational Taxonomic Units (OTU). For each OTU, we selected a representative sequence to assign taxonomy. To assign taxonomy to the remaining quality-controlled sequences, we used the Global Alignment for Sequence Taxonomy (GAST) algorithm

176 (Huse et al., 2008). Our dataset consisted of 1,018,154 sequence reads distributed across

3,750 OTUs.

To compare microbial OTU distributions across samples, we used PRIMER-E statistical software (Clarke and Gorley, 2006) to perform cluster analyses and multidimensional scaling analyses (MDS). Before analysis, we randomly resampled the data so that all samples had the same number of sequence reads (11,078) equivalent to the lowest read sequence number of a single sample (after removing datasets from 5/5 10m and 5/17 10m with <2000 sequences). We also transformed the data to presence/absence counts and created a distance matrix using the Jaccard index of similarity. Using this same randomly resampled and transformed data matrix, we performed a canonical correspondence analysis using CANOCO 4.5 (Microcomputer Power, Ithaca, NY, USA)

(ter Braak and Similauer, 2002) to compare sample microbial communities with environmental geochemical gradients. We used GraphPad Prism version 6 software (La

Jolla, CA, www.graphpad.com) for performing linear regressions and regression comparisons.

2.9 Quantitative polymerase chain reaction

Purified DNA extracts were also subjected to real-time quantitative polymerase chain reaction (qPCR) to gauge haptophyte DNA concentrations with depth and ensure that the sample selected for sequencing was at the point of highest haptophyte cell concentration in the water column. We performed the qPCR reaction using 18S rRNA gene haptophyte specific primers Prym-429F and Prym-887R (Coolen et al., 2004). The qPCR reactions were run on an Applied Biosystems StepOnePlusTM Real-Time PCR System (Foster City,

177 California), using a SYBR Green I assay. Each 20 µl reaction contained 7.2 µl of sterile water, 10 µl of KAPA SYBR® FAST Universal 2X qPCR Master Mix (Woburn, MA),

0.4 µl each of the forward and reverse primers (0.2µM) and 2 µl of template DNA. The qPCR cycling program was after Coolen et al. (2009) and consisted of 38 cycles of denaturation at 94°C for 30s, annealing at 62°C for 40s, primer extension at 72°C for 60s, a photo step of 80°C for 20s. We used between 101 and 106 copies (ten-fold dilution series) of linearized plasmids containing 18S rDNA of Isochrysis galbana CCMP1323 as the external standard to calibrate the copy numbers of haptophyte DNA in the BrayaSø water samples.

3. Results

3.1 Geochemical results

As expected, the Lake George water samples exhibited a gradual warming over the course of the three-month sampling period (Figure 2B) from late April to late July.

Salinity reached a slight plateau on the first half of the summer, and a second peak in the latter half of the seasonal cycle (Figure 2C). Coincident with the second escalation of

salinity we see the decline in C37 alkenone concentrations in the water column (Day 45,

Figure 2A). The C37 alkenone concentrations exhibit a bimodal peak, with a preliminary rise and fall in alkenone concentrations from days 1-22, and again from days 25-70

(Figure 2A).

The three sampling depths generally exhibited similar trends in temperature, salinity and alkenone concentrations. As expected, 10m depth samples had lower temperatures than the upper waters at 0m and 5m depths. Alkenone concentrations

178 peaked at 10m depth slightly delayed after the peak at 0m and 5m depths occurring on

Day 37. C37 alkenone concentrations reached 2.5mg/L at the highest point on Day 37.

Previous to Day 37, nutrient concentrations were low, and at Day 37 reached concentrations where they remained the remainder of our sampling (Figure 3). The lowest concentrations of nitrate, nitrite, and phosphate on Day 24 corresponded to the lowest alkenone concentrations. We will refer to the two distinct peaks in alkenone concentrations as defined by early Peak I from days 1-22 and then later Peak II from days

24-70. Peak I C37 alkenones reached concentrations of 1000ng/L and Peak II reached concentrations of 2500 ng/L (Figure 2A).

k 3.2 In situ alkenone distributions and U37 calibration

The first peak in alkenone concentrations, Peak I, was characterized by both

dominant tetraunsaturated C37:4 alkenones and dominant C37:3 alkenone peaks (Figure 4).

Peak II 5m and 10m depths had a dominant C37:4 alkenone signature that changed to a

dominant C37:3 signature after the maximum alkenone production peak at day 37 (Figure

4). Peak II also saw the appearance of the C38:3 methyl ketones (Figure 4). Peak I alkenones were at their maximum concentration at Day 10 (Figure 2A) and Day 10

alkenone signatures were C37:3 dominant at all depths. Peak II alkenones were at their maximum concentration at depths 0m and 5m at Day 37 (Figure 2A), both characterized

by dominant C37:4. At 10m depth, alkenone concentrations were highest at Day 42 and were also characterized by a dominant C37:4 signature.

k The Lake George in situ U37 calibration was defined by the equation:

179 k U37 = 0.0255T - 0.697

This calibration had an R2 value of 0.712 and a root mean squared error of 3.7°C (Figure

k 5A). The largest scatter in the U37 calibration occurred at about 15°C, corresponding to

Peak II Days 37-45 (Figure 5A). The in situ calibration is most similar to a previously

k reported U37 calibration from Greenland’s lake BrayaSø, a modified in situ calibration

(D’Andrea et al., 2011) that was developed by combining an in situ calibration with a calibration from a suite of German lakes (Zink et al., 2001). Like Lake George, BrayaSø

sediments are also characterized by dominant C37:4 alkenones (Figure 5C).

3.3 In situ haptophyte abundances: qPCR and IT sequencing

All samples between April 26th and June 16th (Days 1-52) were analyzed for haptophyte cell abundances using both qPCR and Ion Torrent Sequencing. This date range spans both Peak I and Peak II, although it does not include the long tail of alkenone concentrations after Peak II. The qPCR results give an idea of total haptophyte cell abundance, as inferred from haptophyte rRNA gene copy numbers. The Ion Torrent results describe the entire eukaryotic microbial community during the haptophyte bloom, and we pay special attention to just the haptophyte Ion Torrent reads.

The qPCR probe targets all haptophyte algae, and like the alkenone data, the qPCR data exhibited a bi-modal distribution (Figure 6). In contrast to the alkenone data,

Peak I had greater haptophyte rRNA gene copy numbers, suggesting greater concentrations of haptophytes cells during Peak I, reaching concentrations of almost 3000 cells/ml in the surface waters on Day 10, the apex of Peak I. As for Peak II, the 0m depth sample reached highest haptophyte cell numbers at Day 42, slightly after the 0m alkenone

180 peak. Curiously, 5m and 10m depth qPCR results did not exhibit the same trends as those in the 0m samples and fluctuated between a few hundred haptophyte cells/ml.

The Ion Torrent data allowed us to more closely examine the haptophyte distributions in Lake George. After OTU clustering, there were two Isochrysidales haptophyte OTUs. After generating V9 haptophyte sequences from cultures of Hap-B

(Table S1), we were able to identify which OTU represented Hap-A and Hap-B

sequences. Hap-A, the inferred dominant C37:4-alkenone producing haptophyte (Chapter

6), and Hap-B, the C38:3 methyl producing haptophyte (Toney et al., 2012) showed bimodal distributions similar to both the alkenone and qPCR data (Figure 4). At 0m depth, Hap-A peaked during alkenone Peak I at Day 10, the same day as the qPCR peak and the alkenone peaks. At 5m and 10m depths, Hap-A was again the more dominant haptophyte during Peak I. At 0m depth, Peak II was dominated by Hap-B OTU sequences, and peaked at Day 42 in accordance with qPCR results, whereas 5m and 10m depth samples peaked at Day 45. At 10m depth, the Hap-A and Hap-B trends were not bimodal, and Hap-A sequences actually peaked at Day 37 followed by Hap-B becoming dominant.

Our randomly resampled Ion Torrent dataset yielded 8172 haptophyte reads across 56 OTUs. If singletons are ignored, the haptophyte yield was 11 OTUs. The haptophyte OTUs group into 5 phyla, Prymnesiales, Phaeocystales, Pavlovales,

Isochrysidales and Coccolithales. The haptophyte phyla distributions are shown in

Figure 7. The Isochrysidales and Pavlovales alternate in abundance, and Prymnesiales is abundant in the early season samples. Coccolithales and Phaeocystales only have one Ion

181 Torrent read each during the seasonal cycle. Of the 1230 Isochrysidales total sequences,

Hap-A and Hap-B comprise 1201 of these sequences (97.6%).

A cluster analysis of the Lake George bloom Ion Torrent sequence data revealed a clear separation between pre- and peak/post-peak bloom samples (Figure 8). The earliest sample, April 26th at 0m depth (LG1_0) was a clear outlier from the other samples, possibly a result of its high Chlamydomonas concentrations (4386 reads, Table S2) and high copepod concentrations (4030 reads, Table S2), a combination unique to only this early surface sample. A canonical correspondence analysis revealed that 70% of the microbial assemblage variation among the samples could be explained by temperature

and salinity (Figure 9). While percent C37:4 was used as an explanatory variable, it is not an environmental variable shaping community composition but instead a variable that clearly fluctuates with the changing microbial community. As expected, Hap-B, the haptophyte that appeared ‘late’ in the bloom event, was closely associated with warmer temperatures as it plots along the temperature axis (Figure 9). There were no clear associations between Hap-A occurrence and environmental variabilities.

k 3.4 Culture studies and U37 calibrations

k Hap-A was grown in culture to determine its U37 calibration and was identified as a picoplankton haptophyte (Chapter 6). Hap-B (Strain #903) was also isolated from Lake

George waters and grown in pure culture at 5,10,15,21 and 24°C. Culture Hap-B alkenones demonstrated the characteristic increase in saturated alkenone isomers with

increasing temperatures (Figure 10A). Hap-B also had a dominant C37:3 alkenone

k signature and the presence of trace levels of C38:3 methyl ketones (Figure 10A). The U37

182 calibration for Hap-B (Figure 5) was distinct from the Hap-A culture at the lower temperatures although their calibrations overlapped at higher temperatures. The Hap-B calibration is defined by the equation

k U37 = 0.0279T – 0.5889 with an R2 value of 0.95563 and an RMSE of 1.7°C. Neither Hap-A nor Hap-B calibrations match that of the in situ Lake George calibration, although the Hap-A

k calibration has a very similar slope to the in situ Lake George U37 calibration. The Hap-B calibration and the Hap-A calibrations are statistically similar and can be explained by a

k single curve, as depicted in Figure 5B. The equation for the composite culture U37 calibration is

k U37 = 0.0326T – 0.671

The Lake George in situ calibration from our study is with an R2 value of 0.92 and

RMSE of 1.7°C. Hap-B displayed a notable change in morphology at different temperatures (Figure 10B), and the highest growth rates were at 15°C and the lowest growth rates at 5°C (Table 1). Alkenone per cell concentrations ranged from 0.011 pg/cell at 15°C to 3.299 pg/cell at 5°C.

3.5 Mock bloom study

To determine if we could recreate the Lake George bloom event in silico, we performed a mock bloom event by incubating Lake George surface sediments and monitoring the supernatant alkenone concentrations and haptophyte DNA over time. Like observed in the environmental samples, the mock bloom event was characterized by Hap-

183 A dominance in the early weeks and the appearance of Hap-B later in the bloom cycle

(Figure 11). Unlike our field samples, the mock-bloom event alkenone signatures were

all C37:4 dominant, although the late-bloom samples had unresolvable alkenone peaks.

The timing of the appearance of Hap-B in the mock bloom event was very similar to that observed in the field, with Hap-B DNA first detected at Day 42 (versus in situ Hap-B rises to dominance in the surface waters at Day 37).

4. Discussion

4.1 Peak I and Peak II haptophyte communities

Alkenone concentrations in the Lake George water column defined Peak I as spanning from Day 1-22 of the sampling effort, and a larger Peak II spanning from Days

24-70. This dual-bloom cycle was confirmed by qPCR analyses, however the qPCR analyses suggested that Peak I was actually a larger, more haptophyte-dense bloom event than Peak II. These two findings are not mutually exclusive, considering alkenone concentrations per cell vary and Peak I may have larger cell numbers with lower alkenone concentrations followed by Peak II with lower haptophyte cell numbers and higher alkenone cell concentrations. Alkenone concentrations vary with cell stage or nutrient concentrations, with cells accumulating alkenones at both lower growth rates

(Popp et al., 1998; Epstein et al., 2001) and low nutrient conditions (Eltgroth et al.,

2005). There may also be genetic predispositions of alkenone production, with different haptophyte strains synthesizing more or less alkenones under identical environmental and physiological conditions (Conte et al., 1998). However, it is important to remember that

184 the qPCR data reflect total haptophyte cell abundances, as opposed to alkenones that are unique to Isochrysidales.

In order to delineate total haptophyte cell numbers from alkenone-producing haptophytes, we refer to the Ion Torrent sequence read abundances. Figure 7 displays the abundance and phylogenetic affiliation of haptophyte OTUs. We can see that like the qPCR results, the Ion Torrent results display a higher number of haptophyte sequences in

Peak I, equaling 5663 sequence tags, and Peak II has only 2506 haptophyte sequence tags over twice as many days. Isochrysidales are almost equal between Peak I and Peak II,

293 and 301 sequence reads, respectively. These results suggest that haptophyte cell abundances are indeed higher during Peak I, as indicated by qPCR results and Ion

Torrent sequencing results, however Isochrysidales cell numbers are almost equal during both peaks. The variation in alkenone concentrations, therefore, between Peak I and Peak

II is likely a reflection of haptophyte taxonomy or cell physiology.

4.2 Environmental bloom conditions

The bimodal alkenone distribution depicts a preliminary haptophyte bloom while nutrient concentrations are low in the spring followed by a second haptophyte bloom coinciding with higher nutrient concentrations in the late spring/early summer. The onset of Peak II coincides with an increase and stabilization of nutrient concentrations in the water column and stable salinity values. Ion Torrent sequencing revealed the dominant haptophyte during Peak I is Hap-A and the dominant haptophyte during Peak II is Hap B.

Our canonical correspondence analysis revealed the association of Hap-A with lower

185 salinities, and the association of Hap-B with higher temperatures and salinities, although this may just be the result of the Hap-B bloom occurring later in the bloom season.

We know from previous research (Chapter 5) that Hap-A is a picoplankter, indicating it may have a competitive advantage in low-nutrient conditions due to its higher surface-area to volume ratio. The early season Peak I water column, when the seasonal ice melt is still mixing throughout the lake’s water column, may allow picoplanktonic Hap-A to outcompete larger, and more nutrient-dependent Hap-B. While temperature and salinity explained 70% of the total eukaryotic OTU variance, it remains to be seen if warmer temperatures and higher salinities are requirements for a Hap-B bloom.

4.3 Hap-A and Hap-B alkenone production

As suggested in a previous study, and confirmed in this one, Hap-B is a dominant

C37:3 alkenone producer, similar to its close relatives Pseudoisochrysis paradoxa and

Chrysotila lamellosa (Toney et al., 2012; Theroux et al., 2010). DNA sequences placed

Hap-A as a close relative of an uncultured haptophyte from Ace Lake, Antarctica (Coolen

et al., 2004), also a presumed dominant C37:4 alkenone producers. Our mock-bloom results suggest that when Hap-A is dominant, the alkenone signature is C37:4 dominant

(Figure 11). Hap-B, in contrast, produced only C37:3-dominant signatures in culture

(Table 1). Our in situ alkenone lipid signatures from Lake George did not reflect this clear separation between Hap-A dominance and Hap-B dominance—instead, during both

Peak I and Peak II there were alternating C37:4- and C37:3-dominant alkenones. The culture studies indicate that Hap-A and Hap-B do have distinct alkenone lipid signatures at lower

186 temperatures, with Hap-A producing abundant C37:4 signatures and Hap-B producing only

C37:3-dominant alkenone signatures.

As suggested from previous culture studies, Hap-A may lose its C37:4 dominant alkenone signature when at higher growth rates as experienced in pure culture (Chapter

6) and at certain times in the bloom cycle. If we examine the Ion Torrent Hap-A and Hap-

B abundances more closely (Figure 4), we see this pattern exhibited in the 0m depth samples. When Hap-A is at low growth rates early in the season, the alkenone signature

is characterized by an abundant C37:4 alkenone. When Hap-A reaches it peak cell

concentrations, at Day 10, the alkenone signature is characterized by dominant C38:3 alkenones. Again, when Hap-A concentrations drop at Day 22, the alkenone signature is

C37:4 dominant. This pattern is not maintained when Hap-B appears in the water column, at which point there is a mixture of competing alkenone fingerprints as indicated by

abundant C37:4 alkenones co-occurring with C38:3 methyl ketones.

k 4.4 U37 alkenone unsaturation indices

Genotypic effects on alkenone production are evident in the individual alkenone

k signatures for Hap-A and Hap-B cultures. However, Hap-A and Hap-B U37 calibrations

k could be explained by a single composite calibration (Figure 5B). The higher U37 values derived from haptophytes cultures as opposed to the in situ measurements may be the result of artificially high growth rates experienced by these cells in culture. The in situ

k U37 calibration for Lake George is closely aligned with previous calibrations from lake

BrayaSø, Greenland (D’Andrea et al., 2011) and from a series of German lakes (Zink et al., 2001).

187 The Hap-A and Hap-B cultures had similar growth rates, with doubling times average 3-4 days, with the exception of the 5°C Hap-B culture with very low growth

k rates and very high alkenone concentrations (Table 1). The cultures have similar U37 values for 15°C, which is curious considering this is the temperature at which the in situ

k U37 calibration has the highest degree of scatter (Figure 5A). This may indicate the samples collected on the 15°C days, which center around Days 37-55, may have been during high wind, high mixing conditions leading to transport of alkenones throughout

k the water column and erroneous U37 estimations. Indeed, Days 45 and 52 have overlapping temperature measurements for 0m and 5m depths, indicating mixing of the surface waters. Future in situ temperature estimates should use caution when interpreting alkenone concentrations from days with low stratification indices.

Both Ion Torrent sequencing and qPCR results showed much lower Isochrysidales cell concentrations at 5m and 10m depths in the water column. Ion Torrent Isochrysidales sequences averaged 48 reads in the 0m samples, 17 reads at 5m depth, and 12 reads at

k 10m depth. We recalculated the U37 for each individual depth (Figure S1) and 0m, 5m,

k and 10m samples all had an increasing U37 with increasing temperature, suggesting alkenones collected at depth still record a temperature response. Both qPCR and Ion

Torrent data indicate a higher haptophyte cell concentration at 0m depth, and alkenone concentrations grade slightly from 0m > 5m > 10m, suggesting the majority of alkenones being produced in the Lake George water column are from 0m depth.

4.5 Implications for paleothermometry

188 The majority of Lake George downcore sediments are C37:4 dominant, with a few exceptions when alkenone signatures have dominant C37:3 and the presence of the C38:3 methyl ketone, a hallmark of the Hap-B cultures (Toney, 2011; Toney et al., 2012).

These C37:3 dominant signatures may be the result of environmental conditions favoring

Hap-B, such as warmer water temperatures, higher salinities or higher nutrient concentrations as suggested by our field observations.

Neither the Hap-A nor the Hap-B ex situ calibrations yielded an exact match to the Lake George in situ calibration, likely because pure culture conditions do not accurately replicate those experienced in the field. Nutrient cycling, variable temperature regimes, and predator/prey fluctuations are not captured in pure culture studies, and may be impacting the in situ alkenone signatures. With this in mind, we believe the in situ calibration most accurately captures the haptophyte response to water temperature in

Lake George and the in situ calibration would be the most reliable metric for reconstructing temperatures from the Lake George alkenone record.

5. Conclusions

Our study of a seasonal haptophyte bloom in Lake George, ND provides new insights into lacustrine haptophyte physiology, ecology and bloom conditions.

Throughout a seasonal cycle, the two species of haptophytes in Lake George alternate in abundance, as first indicated by in situ alkenone concentrations. DNA sequencing revealed that the bimodal alkenone distribution reflected alternating abundances of species Hap-A and Hap-B. Hap-A was dominant in the early bloom stage Peak I, and

Hap-B was dominant during subsequent Peak II. As evidenced by both field observations

189 and our Hap-B culture manipulations, we have confirmed that Hap-B is responsible for

dominant C37:3 and C38:3 methyl ketone production and Hap-A is the dominant tetraunsaturated alkenone producer. The dual peak in alkenone concentrations throughout the course of the seasonal cycle was again reflected in qPCR haptophyte gene copy numbers and the Hap-A to Hap-B succession was also observed in our mock-bloom enrichment culture experiment. These complimentary results demonstrate how our multi- disciplinary approach enabled a greater resolution of haptophyte cell abundances than was previously possible through geochemical analyses alone.

The environmental triggers for Hap-A and Hap-B dominance are still unclear.

While Hap-B appears during warmer temperatures and higher salinities, it remains to be seen if these conditions are a prerequisite for its bloom. Efforts are currently underway to determine if, in a mock-bloom experiment, a warmer temperature incubation results in a preliminary bloom by Hap-B instead of Hap-A. Clearly, the interplay between the length of Peak I and Peak II, and the abundance of individual Hap-A and Hap-B, results in variable inputs of alkenones into the sediment record. It remains to be seen if Hap-A alkenones are preferentially exported to the Lake George sediments, given the

characteristic C37:4 dominant peak in the sediment record. Although futher research is

k needed, we believe the in situ U37 calibration provides an accurate representation of the relationship between alkenone-unsaturation and lake water temperature and therefor provides a means of reconstructing temperature from the Lake George sediment record.

Acknowledgements

190 This work was supported by a National Science Foundation award to Y. Huang (EAR-

1122749) and L. Amaral-Zettler (EAR-1124192), a Brown SEED fund to Y. Huang and

L. Amaral-Zettler, and an American Association of University Women dissertation fellowship to S. Theroux.

191

References

Amaral-Zettler LA, McCliment EA, Ducklow HW, Huse SM. (2009). A method for studying protistan diversity using massively parallel sequencing of V9 hypervariable regions of small-subunit ribosomal RNA genes. PLoS One 4:e6372.

Brassell SC, Eglinton G, Marlowe IT, Pflaumann U, Sarnthein M. (1986). Molecular stratigraphy - a new tool for climatic assessment. Nature 320:129-133.

Chu GQ, Sun Q, Li SQ, Zheng MP, Jia XX, Lu CF, Liu JQ, Liu TS. (2005). Long-chain alkenone distributions and temperature dependence in lacustrine surface sediments from China. Geochimica Et Cosmochimica Acta 69:4985-5003.

Clarke KR, Gorley RN. (2006). PRIMER v6: User Manual/Tutorial.

Conte MH, Thompson A, Lesley D, Harris RP. (1998). Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica. Geochimica Et Cosmochimica Acta 62:51-68.

Coolen MJL, Muyzer G, Rijpstra WIC, Schouten S, Volkman JK, Sinninghe Damsté J. (2004). Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth and Planetary Science Letters 223:225-239.

Coolen MJL, Saenz JP, Giosan L, Trowbridge NY, Dimitrov P, Dimitrov D, Eglinton TI. (2009). DNA and lipid molecular stratigraphic records of haptophyte succession in the Black Sea during the Holocene. Earth and Planetary Science Letters 284:610- 621.

D'Andrea WJ, Huang Y. (2005). Long chain alkenones in Greenland lake sediments: Low δ13C values and exceptional abundance. Organic Geochemistry 36:1234-1241.

D'Andrea WJ, Lage M, Martiny JBH, Laatsch AD, Amaral-Zettler LA, Sogin ML, Huang YS. (2006). Alkenone producers inferred from well-preserved 18S rDNA in Greenland lake sediments. Journal of Geophysical Research-Biogeosciences 111:doi:10.1029/2005JG000121.

D'Andrea WJ, Huang Y, Fritz SC, Anderson NJ. (2011). Abrupt Holocene climate change as an important factor for human migration in West Greenland. Proceedings of the National Academy of Sciences

Eltgroth ML, Watwood RL, Wolfe GV. (2005). Production and cellular localization of neutral long-chain lipids in the haptophyte algae Isochrysis galbana and Emiliania huxleyi. Journal of Phycology 41:1000-1009.

192

Epstein BL, D'Hondt S, Hargraves PE. (2001). The possible metabolic role of C-37 alkenones in Emiliania huxleyi. Organic Geochemistry 32:867-875.

Fritz SC. (2011). Lake George and its geological history during the last ~10,000 years. CGREC Annual Report

Guillard RRL (1975) Culture of phytoplankton for feeding marine invertebrates. In Culture of marine invertebrate animals. WL Smith, MH Chanley, eds. Plenum Press, New York, p. 26.

Huse SM, Mark Welch D, Morrison HG, Sogin ML. (2010). Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environmental Microbiology 1889-1898.

Huse SM, Huber JA, Morrison HG, Sogin ML, Welch DM. (2007). Accuracy and quality of massively parallel DNA pyrosequencing. Genome biology 8:R143.

Huse SM, Huber JA, Welch DM, Relman DA, Sogin ML. (2008). Exploring microbial diversity and taxonomy using SSU rRNA hypervariable tag sequencing. PLoS genetics 4:e1000255.

Liu WG, Liu ZH, Fu MY, An ZH. (2008). Distribution of the C37 tetra-unsaturated alkenone in Lake Qinghai, China: A potential lake salinity indicator. Geochimica Et Cosmochimica Acta 72:988-997.

Müller PJ, Kirst G, Ruhland G, von Storch I, Rosell-Mele A. (1998). Calibration of the k' alkenone paleotemperature index U37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S). Geochimica Et Cosmochimica Acta 62:1757-1772.

Pearson EJ, Juggins S, Farrimond P. (2008). Distribution and significance of long-chain alkenones as salinity and temperature indicators in Spanish saline lake sediments. Geochimica Et Cosmochimica Acta 72:4035-4046.

Popp BN, Kenig F, Wakeham SG, Laws EA, Bidigare RR. (1998). Does Growth Rate Affect Ketone Unsaturation and Intracellular Carbon Isotopic Variability in Emiliania Huxleyi? Paleoceanography 13:35-41.

Prahl FG, Wakeham SG. (1987). Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330:367-369.

Rontani JF, Prahl FG, Volkman JK. (2006). Re-examination of the double bond positions in alkenones and derivatives: biosynthetic implications. Journal of Phycology 42:800-813.

193 k' Sikes EL, Volkman JK. (1993). Calibration of alkenone unsaturation ratios (U37 ) for paleotemperature estimation in cold polar waters. Geochimica et Cosmochimica Acta 57:1883-1889.

Simon N, Campbell L, Örnolfsdȯ ttiṙ E, Groben R, Guillou L, Lange M, Medlin LK. (2000). Oligonucleotide probes for the identification of three algal groups by dot blot and fluorescent whole-cell hybridization. The Journal of Eukaryotic Microbiology 47:76-84.

Sun Q, Chu GQ, Liu GX, Li S, Wang XH. (2007). Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellosa (Haptophyceae). Organic Geochemistry 38:1226-1234.

Ter Braak CJF, Smilauer P. (2002). CANOCO reference manual and CanoDraw for Windows user's guide: Software for canonical community ordination (version 4.5). Microcomputer Power, Ithaca, NY.

Theroux SM, D'Andrea WJ, Toney JL, Amaral-Zettler LA, Huang Y. (2010). Phylogenetic diversity and evolutionary relatedness of alkenone producing haptophyte algae in lakes: Implications for continental paleotemperature reconstructions. Earth and Planetary Science Letters 300:311-320.

Toney JL, Huang Y, Fritz SC, Baker PA, Grimm E, Nyren P. (2010). Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes in the interior United States. Geochimica Et Cosmochimica Acta

Toney, JL. (2011). Climatic and environmental controls on the occurrence and distributions of long-chain alkenones in lakes of the interior United States. PhD thesis, Brown University. 351pgs.

Toney JL, Theroux S, Andersen RA, Coleman A, Amaral-Zettle L, Huang Y. (2012). Culturing of the first 37:4 predominant lacustrine haptophyte: Geochemical, biochemical, and genetic implications. Geochimica et Cosmochimica Acta 78:51 - 64.

Tyrrell T, A Merico (2004). Emiliania huxleyi: bloom observations and the conditions that induce them In Coccolithophores-from molecular processes to global impact. HR Thierstein, J Young, eds. Springer, p. 75.

Whitehead RL. (1996). Groundwater atlas of the United States - Segment 8. USGS

Zink KG, Leythaeuser D, Melkonian M, Schwark L. (2001). Temperature dependency of long-chain alkenone distributions in Recent to fossil limnic sediments and in lake waters. Geochimica Et Cosmochimica Acta 65:253-265.

194 Table 1. Haptophyte Hap-B #903 culture cell counts and alkenone concentrations.

Growth Final cell rate concentration (Doubling C37:4 C37:3 C37:2 C37total C37alk/cell Temp. (°C) (cell/ml) time) (ng/L) (ng/L) (ng/L) (ng/L) % C37:4 (pg) Uk37 Uk'37 5 10000 71.44 16363 16375 251 32989 49.6 3.299 0.02 -0.49 10 640000 3.64 8831 13779 1515 24125 36.6 0.038 0.10 -0.30 15 1600000 3.009 3583 12855 1951 18390 19.5 0.011 0.13 -0.09 21 250000 4.632 1861 15067 884 17812 10.4 0.071 0.06 -0.05 24 310000 4.359 367 2649 640 3656 10.0 0.012 0.19 0.07

195

Figure 1. Map of North Dakota (inset) and Lake George, ND.

196 3000! Peak I Peak II 2500!

2000! !

1500! 0m! 5m! 1000! 10m! C37 total (ng/L) C37 total 500!

0! 0! 5! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Day 1 = April 26, 2011!

30!

25!

! 20!

15! 0m! 10! 5m! 10m!

Temperature (°C) Temperature 5!

0! 0! 5! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Day 1 = April 26, 2011! 14! 12! 10! ! 8! 5m! 6! 10m!

Salinity (ppt) Salinity 4! 0m! 2! 0! 0! 5! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Day 1 = April 26, 2011!

Figure 2. Lake George water column throughout the sampling period at 0m, 5m,

and 10m depths. A. C37-alkenone concentrations. B. Temperature C. Salinity. Peaks

I and II as indicated.

197

1000! 0m! 10m! 5m! ! 100! ug/L 2

NO 10!

1!

1000! ! 100! ug/L 3

NO 10!

1! 1000! ! 100! ug/L 4 10! PO

1! 1! 10! 22! 24! 37! 42! 45! 52! 72! 77! 88! 94! Day 1 = April 26, 2011!

Figure 3. Nutrient concentrations during the Lake George seasonal cycle.

198

Figure 4. Percent concentrations of alkenone isomers and Ion Torrent sequence read numbers of Hap-A and Hap-B throughout the seasonal cycle at 0m, 5m, and

10m depths. Alkenone concentration Peaks I and II as indicated. Asterisk denotes a

sample with C38:3 methyl ketones present.

199 A. 0.2! 0.1! Hap-A! y = 0.0373x - 0.7531! Hap-B! 0.0! R" = 0.94! y = 0.0279x - 0.5889! R" = 0.94! -0.1! ! -0.2!

Uk37 -0.3! -0.4! In situ calibration ! -0.5! Uk37 = 0.0255T - 0.697! R" = 0.71! -0.6! -0.7! 0! 5! 10! 15! 20! 25! 30! Temperature (°C)! B. 0.2! Culture Uk37! 0.1! y = 0.0326x - 0.671! 0.0! R" = 0.92! -0.1! ! -0.2!

Uk37 -0.3! -0.4! In situ calibration! -0.5! Uk37 = 0.0255T - 0.697! R" = 0.71235! -0.6! -0.7! 0! 5! 10! 15! 20! 25! 30! Temperature (°C)!

C. 1! German lakes! Lake BrayaSø! 0.8! E. huxleyi! P. paradoxa! 0.6! Polar waters! 0.4! C. lamellosa! LG2011! 0.2! ! 0! Uk37 -0.2!

-0.4!

-0.6!

-0.8!

-1! 0! 5! 10! 15! 20! 25! Temperature (°C)!

k k Figure 5. U37 calibrations for the Lake George field and culture alkenones. A. U37 calibrations for Lake George 2011 in situ calibration, Hap-B isolate #903, and Hap-

200 k A <3µm culture. In situ calibration 2011 from this work. B. Composite “culture U37

” calibration from Hap-A and Hap-B calibrations plotted with in situ calibration. C.

k Lake George (LG2011) in situ calibration compared to previously published U37 calibrations. References are as follows: German Lakes, Zink et al., 2001; Lake

George in situ, Toney et al., 2010; Lake BrayaSø, D'Andrea et al., 2011; E. huxleyi,

Prahl et al., 1988; P. paradoxa, Chapter 4; Polar waters, Sikes and Volkman 1993;

Lake George in situ 2011, Theroux, this work; C. lamellosa, Sun et al., 2007.

201 0m! 5m! 10m! 0m! 5m! 10m! ! 5000! 3000!

4500! 2000!

4000! 1000! 0! 3500! -1000! ! 3000! -2000! 2500! -3000! Alkenone (ng/L) concentrations 2000! -4000! 1500! -5000! 1000! -6000! 500! -7000! 0! -8000! 0! 5! 10! 15! 20! 25! 30! 35! 40! 45! 50! 55! 60!

Haptophyte 18S rRNA gene copies/ml gene Haptophyte 18S rRNA Day 1 = April 26, 2011!

Figure 6. qPCR haptophyte gene copy number concentrations throughout the course of the Lake George seasonal cycle. Alkenone concentrations as a comparison.

202 ! 0m! 500! Peak I! Peak II! 400! Coccolithales! 300! Isochrysidales! 200! 100! Pavlovales! 0! Phaeocystales! 1! 10! 22! 24! 37! 42! 45! 52! Ion Torrent sequence reads Torrent Ion Prymnesiales! Day 1 = April 26, 2011!

! 5m! 2000!

1500! Coccolithales! 1000! Isochrysidales! 500! Pavlovales! 0! Phaeocystales! 1! 10! 22! 24! 37! 42! 45! 52! Prymnesiales! Ion Torrent sequence reads Torrent Ion Day 1 = April 26, 2011!

! 10m! 4000!

3000! Coccolithales! 2000! Isochrysidales! 1000! Pavlovales! 0! Phaeocystales! 1! 24! 37! 42! 45! 52! Prymnesiales! Ion Torrent sequence reads Torrent Ion Day 1 = April 26, 2011!

Figure 7. Haptophyte phyla abundances from Ion Torrent OTUs at 0m, 5m, and

10m depths. Peak I and Peak II as indicated. OTUs are defined at 97% similarity and the dataset was randomly resampled to yield equal OTU numbers per sample.

203 A .

B.

Figure 8. A. Cluster analysis of Lake George Ion Torrent microbial community similarity among samples. All sample OTU abundances were first transformed into presence/absence data and then Bray-Curtis similarities were computed before analysis. Bloom stage is overlain and defined as Pre-bloom (Days 1-22), Peak (Days

24-42), Post (Day 45-52). B. Multi-dimensional scaling analysis as in (A). All sample

OTU abundances were first transformed into presence/absence data and then

204 Jaccard similarities were computed before analysis. Bloom stage is overlain and defined as Pre-bloom (Days 1-22), Peak (Days 24-42), Post (Day 45-52).

205

Figure 9. Canonical correspondence analysis of Lake George geochemical gradients and OTU abundances. All OTUs were randomly resampled, transformed to presence/absence and Jaccard distances were computed before analysis. Hap-A and

Hap-B are indicated on the graph. Fist and second axis are defined by temperature

and percent C37:4, respectively, explaining 70% of the OTU variance. Salinity is the third explanatory variable and the rest of the geochemical data are supplemental variables. Peak I is defined as pre-bloom, and Peak II by peak-bloom and post- bloom.

206 Hap-B #903 C C A 38:3et! Hap-B #903 38:3et! C 5°C! 21°C! 37:3! ! ) pA

C38:4! C38:2! Response ( C37:3! C C37:4!

! 37:4! C 38:4! ! C ! 37:2! ! C38:2! !

! C37:2! !

Retention time (min)! Retention time (min)!

B . 5°C 21°C

10µ 10µ

Figure 10. Hap-B strain #903 alkenones and morphology. A. Gas chromatogram of

Hap-B #903 alkenone signature at 5°C and 21°C. B. Hap-B cells. Hap-B is shown at

5°C and 21°C to illustrate the cell morphology change with temperature.

207

700! 100!

90! ! ! 600! 80! g/L) " 500! 70! 60! 400! C37:4! 50! C37:3! 300! 40! C37:2! %Hap-A! 200! 30! %Hap-B! 20! Alkenone concentraiton ( Alkenone concentraiton

100! Haptophyte OTU abundance (%) 10! 0! 0! 9! 11! 14! 17! 21! 24! 28! 31! 35! 38! 42! 45! Enrichment culture age (days)!

Figure 11. Mock bloom event alkenone signatures and haptophyte DNA sequence taxonomy.

208 Supplemental

Table S1. Hap-B culture 18S rRNA partial gene sequence. Generated using primers

429-Forward/EukB-Reverse.

>HapB GCGCGTAAATTGCCCGAATCCTGACACATTGAGGTAGTGACAAGAAATAA CAATACAGGGCTCTTCGAGTCTTGTAATTGGAATGAGTACAATTTACATC TCTTCACGAGGATCAATTGGAGGGCAAGTCTGGTGCCAGCAGCCGCGGTA ATTCCAGCTCCAATAGCGTATACTAAAGTTGTTGCAGTTAAAACGCTCGT AGTCGGATTTCGGGGCGGGCCCGCCGGTCTGCCGATGGGTATGCACTGGC GGGCGCGTCCTTCCTCCCGGAGACGGCTGCTACTCCTAACTGAGCGGTGG CCGGAGACGGGATATTTACTTTGAAAAAATCAGAGTGTTTCAAGCAGGCA GTCGCTCTTGCATGGATTAGCATGGGATAATGAAATAGGACTCTGGTGCT ATTTTGTTGGTTTCGAGCACCGGAGTAATGATTAACAGGGACAGTCAGGG GCACTCGTATTCCGCCGAGAGAGGTGAAATTCTCAGACCAGCGGAAGACG AACCACTGCGAAAGCATTTGCCAGGGATGTTTTCACTGATCAAGAACGAA AGTTAGGGGATCGAAGACGATCAGATACCGTCGTAGTCTTAACCATAAAC CATGCCGACTAGGGATTGGAGGATGTTCCATTTGTGACTCCTTCAGCACC TTTCGGGAAACTAAAGTCTTTGGGTTCCGGGGGGAGTATGGTCGCAAGGC TGAAACTTAAAGGAATTGACGGAAGGGCACCACCAGGAGTGGAGCCTGCG GCTTAATTTGACTCAACACGGGGAAACTTACCAGGTCCAGACATTGTGAG GATTGACAGATTGAGAGCTCTTTCTTGATTCGATGGGTGGTGGTGCATGG CCGTTCTTAGTTGGTGGAGTGATTTGTCTGGTTAATTCCGTTAACGAACG AGACCGCAGCCTGCTAAATAGCGTGCCGAACCCTTTGTTGGGCGTCGCTT CTTAGAGGGACAACTTGTCTTCAACAAGTGGAAGTTCGCGGCAATAACAG GTCTGTGATGCCCTTAGATGTTCTGGGCCGCACGCGCGCTACACTGATGC ATTCAGCGAGTCTATCACCTTGACCGAGAGGTCCGGGTAATCTTTTGAAA TTGCATCGTGATGGGGATAGATTATTGCAACTATTAATCTTCAACGAGGA ATTCCTAGTAAGCGTGTGTCATCAGCGCACGTTGATTACGTCCCTGCCCT TTGTACACACCGCCCGTCGCTCCTACCGATTGAATGATCCGGTGAGGCCC CCGGACTGCGGCTCCGTCGCTGGTTCTCCAACTTTGGGGTTGCGGGAAGC TGTCCGAACCTTATCATTTAGAGGAAGGAGAAGTCGTAACAAGGTTTCCG TAGGTGAACCTGCAGAGGATCA

209 Table S2. Most abundant eukaryote OTU Ion Torrent sequence reads. Hap-A and

Hap-B abundances are included. OTUs were defined at 97% similarity. Raw data was randomly resampled to yield equal sequence read count for each sample

(11,078). Peak I and Peak II as indicated.

210

211

k Figure S1. U37 calibration using only samples from 0m, 5m, and 10m depths.

212

CHAPTER 8

CONCLUSIONS

SUSANNA M. THEROUX

BROWN UNIVERSITY, Department of Geological Sciences MARINE BIOLOGICAL LABORATORY, Josephine Bay Paul Center

213 1. Conclusions

Haptophyte algae have supplied paleoclimatologists with one of the most powerful methods to reconstruct ancient climate history. The discovery of novel alkenone-producing haptophyte species provides new challenges for this 25-year-old proxy. However, the allure of accurate, alkenone-based, continental paleotemperature reconstructions makes the endeavor worthwhile.

This thesis employs a heavily biological approach to understand the genetic and physiological determinants of alkenone sedimentary records. We have expanded the known species diversity of lacustrine alkenone-producing haptophytes and found that haptophyte species can have worldwide distributions, with single species being found in both lakes at the Arctic Circle and the Tibetan Plateau. We discovered that sedimentary records can be composites of multiple haptophyte species and that it is not always possible to decipher the haptophyte species identity based upon an alkenone fingerprint.

We have also found that fluctuating concentrations of C37:4 alkenones are likely a reflection of haptophyte community changes, as opposed to physiological changes by a single haptophyte species.

k The isolation of novel haptophyte species enabled us to compare in situ U37 estimations with ex situ calibrations. We have now expanded the dataset of cultured

k haptophyte species with lab-based U37 calibrations, adding Pseudoisochrysis paradoxa and Lake George Hap-B to the list. Our isolation of Hap-A has been an achievement in

endurance, and the discovery that this species it is a picoplankter and a dominant C37:4 producer suggests it is specifically adapted for long winters and oligotrophic waters.

214 The monitoring of the Lake George haptophyte bloom event refined our understanding of alkenone production in Lake George and alkenone production by two novel haptophyte species. As suspected, Hap-A was present during low nutrient

conditions in the early bloom season, producing predominantly C37:4 alkenones and Hap-

B was present in the late bloom season when nutrient concentrations were higher and waters were more stable. We were able to replicate this bloom event in an enrichment culture, monitoring haptophyte species over time, and just as observed in the field, Hap-A is the dominant species during the early stages of the bloom. Our culturing results

k determined the Hap-A and Hap-B U37 calibrations are statistically identical but distinct from the in situ calibration, and we believe the in situ calibration reflects the most accurate relationship between water temperature and alkenone unsaturation in Lake

George.

At times, DNA sequencing makes the haptophyte community, and therefore the alkenone record, appear dauntingly complex. However, DNA sequencing is able to resolve complexities in haptophyte communities that are otherwise indecipherable or overlooked. This thesis stands in support of a multi-disciplinary approach to alkenone-

k based paleothermometry; the U37 index will be better for it.

215